(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Biodiversity Heritage Library | Children's Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Geographic variation and evolutionary relationships among broad-clawed shrews of the Cryptotis goldmani-group (Mammalia, Insectivora, Soricidae)"

UNIVERSITY OF 
ILLINOIS LIBRARY 

AT URBAiMA-CHAMPAIGN 

BIOLOGY 

MP 



DDDD 


> CO 

o > 


O — 1 O CO 


co o 


c: n: cz 3 




co m co ^ 




ZJ CO -1 2 

m co 2 o 


o 

< 



2 F* 



DDDD 



h r c 

A H 3 
o tr 

qd rj 

U 

• ^ 



H Q CO 


C* 



a. o 



DD 



□□DDDD 



U^ Ifl H 
Hd Oh- 



■4 3 
H 



*] be 



o 
o 
o 

•1 



l\3 



CO 


»r. 


til 


IT. 


a 


3 


T 


M, 


u 


■ 


_ 


(O 


< 


U 




• ■ 


CD 


4 




I 


U.' 






Ul 


ui 








tn 









CTi 



O in 
u> 

CD CD 

CD I 

I to 

li.i ID 

cn 



"ST 

8 

f 



J 



M 



r» 
n 
H 

0. 

H 
■fl- 



ID U CO 







TU 

ridfiici 



H 



CO 

3- 
H- 

1 

II 
3 
(t 




■£" 



I- 

QD 





Zoology 



590 .5 

FI 

N.S. 



navi 



J* 1 ?" 1 "' 1 -J-SlQ 



BIX 



graphic Variation and wy 

Ktionships Among Broad-Clawed 
ws of the Cryptoti dmani-G roup 
immalia: Insectivora: Soricidat 






& 



999 

197 



T : i > 



LISHED BY F 



. 



I 



i^WACserr:* m 



FIELDIANA 



Zoology 

JEW SERIES, NO. 91 






eographic Variation and Evolutionary 
Relationships Among Broad-Clawed 
Shrews of the Cryptotis goldmani-Group 
Mammalia: Insectivora: Soricidae) 



s ea I Woodman 

lepartment of Biological Sciences 

ast Stroudsburg University 

ast Stroudsburg, Pennsylvania 18301 



Robert M. Timm 

Natural History Museum and 
Department of Systematics and Ecology 
University of Kansas 
awrence, Kansas 66045-2454 



accepted August 7, 1997 
Published January 13, 1999 
Publication 1497 



PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY 



© 1999 Field Museum of Natural History 

ISSN 0015-0754 

PRINTED IN THE UNITED STATES OF AMERICA 



Table of Contents 

Abstract 1 

Resumen 1 

Introduction 2 

Methods 3 

Systematic Biology 5 

"Cryptotis mexicana-group" Choate, 1970 .... 5 

"Cryptotis goldmani-group" 5 

Cryptotis goodwini Jackson, 1933 8 

Cryptotis goodwini goodwini Jackson, 

1933 8 

Cryptotis goodwini magnimana, new 

subspecies 11 

Cryptotis griseoventris Jackson, 1933 ... 16 

Cryptotis goldmani (Merriam, 1895) 19 

Cryptotis alticola (Merriam, 1895) 21 

Phylogeny of the Cryptotis goldmani- 

Group 23 

Evolution of the Forelimb 25 

Zoogeography of the Cryptotis goldmani- 

Group 30 

Key to the Cryptotis coldm an /-Group 31 

Acknowledgments 31 

Literature Cited 32 

Appendix I: Transition Series Used in 

Phylogenetic Analysis 33 

Appendix II: Additional Specimens Exam- 
ined 34 



List of Illustrations 

1. Left humerus of Cryptotis goldmani, 
indicating anatomical features men- 
tioned in the text 3 

2. Measurements of the skull used in this 
study 4 

3. Lateral view of the left zygomatic 
plates of Cryptotis parva orophila, C. 
nigrescens, and C. mexicana 7 

4. Map of southern Mexico and north- 
western Central America showing geo- 
graphic distribution of the Cryptotis 
goldmani- group 9 

5. Three-dimensional plot showing simi- 
larities and differences among speci- 
mens of Cryptotis goodwini from Gua- 
temala and El Salvador 12 

6. Dorsal and ventral views of the crani- 
um and lateral view of the skull of the 
holotype of Cryptotis goodwini magni- 
mana, new subspecies 13 



7. Plot of the first two factors from prin- 
cipal components analysis of speci- 
mens of Cryptotis griseoventris, C. 
goodwini goodwini, and C g. magni- 
mana, new subspecies 14 

8. Plot of size against geographic locality 
for specimens of Cryptotis griseoven- 
tris, C goodwini goodwini, and C. 
goodwini magnimana, new subspecies ... 16 

9. Plot of length of unicuspid toothrow 
against condylobasal length for Crypto- 
tis goodwini goodwini, C. goodwini 
magnimana, new subspecies, and C. 
griseoventris 17 

10. Plot of breadth of zygomatic plate 
against length of palate for Cryptotis 
goodwini goodwini, C. goodwini mag- 
nimana, new subspecies, and C. gri- 
seoventris 17 

1 1 . Plot of width of palate (M2B) against 
condylobasal length for Cryptotis altico- 
la, C goldmani, and C. griseoventris 18 

12. Lateral views of the orbital areas of 
the skulls of Cryptotis griseoventris 

and C. goldmani 20 

13. Topologies of the nine shortest trees 
from phylogenetic analysis of the 
Cryptotis mexicana-group using 29 
morphological transition series 24 

14. Topology of the strict consensus tree of 
the Cryptotis mexicana-group con- 
structed from the nine shortest-length 

trees 26 

1 5 . Left humeri of selected Cryptotis 27 

16. Ventral view of the right forefoot of 
Cryptotis parva orophila, C. peregrina, 

C. mexicana, and C. goldmani 28 

17. Dorsal views of bones of the left ma- 
nus of Cryptotis parva orophila, C. 
peregrina, C mexicana, and C. gold- 
mani 28 

18. Dorsal views of bones of the left pes 
of Cryptotis parva orophila, C. nigres- 
cens, C. peregrina, C mexicana, C. al- 
ticola, and C. goldmani 29 



List of Tables 



1 . Measurements of Cryptotis used in this 
study 6 

2. Comparisons of characters among taxa in 

the Cryptotis goldmani-group 8 



3. Factor loadings for the first two factor 
axes from principal components analysis 
of Cryptotis griseoventris, C. goodwini 
goodwini, and C. goodwini magnimana, 

new subspecies 14 

4. Correlation matrix for the eight variables 
used in principal components analysis of 
Cryptotis griseoventris, C. goodwini 
goodwini, and C. goodwini magnimana, 

new subspecies 15 

5. Latitude and longitude for collecting sites 



of Cryptotis griseoventris, C goodwini 
goodwini, and C. goodwini magnimana, 

new subspecies 

Factor loadings for the first three factor 
axes from principal components analysis 
of Cryptotis goodwini from Guatemala 

and El Salvador 

Complete character matrix for phyloge- 
netic analysis of members of the Crypto- 
tis mexicana-group and two outgroups, 
Cryptotis parva parva and Cryptotis ni- 
grescens 2 



Geographic Variation and Evolutionary Relationships 
4mong Broad-Clawed Shrews of the Cryptotis 
roldmani-Group (Mammalia: Insectivora: Soricidae) 



Veal Woodman and Robert M. Timm 



Abstract 

The Cryptotis goldmani-group of small-eared shrews consists of species that occupy high- 
elevation (>1000 m) habitats in Mexico and northern Central America. Previously, this group 
was viewed as consisting of only two species (C. goldmani and Cryptotis goodwini) that were 
characterized by extreme enlargement of the forefeet and foreclaws. Phylogenetically, C. gold- 
mani and C goodwini were placed in Choate's (1970) Cryptotis mexicana- group, which oth- 
erwise consisted of the four subspecies of C. mexicana. Our reevaluation of these shrews 
indicates that the subspecies of C. mexicana are well-differentiated taxa that we recognize as 
separate species (C. mexicana, Cryptotis nelsoni, Cryptotis obscura, and Cryptotis peregrina). 
The C. goldmani-group consists of at least four distinct species: Cryptotis alticola, in Colima, 
Jalisco, Mexico, Michoacan, Puebla, and the Distrito Federal of Mexico; C. goldmani, in the 
Sierra Madre del Sur of Guerrero and Oaxaca; C. goodwini, in Chiapas, Guatemala, El Salvador, 
and Honduras; and Cryptotis griseoventris, in the northern highlands of Chiapas and Guatemala. 
A single new specimen of C. goodwini indicates the presence of this species in Honduras, 
which is outside of the previously recognized geographic range of the C. mexicana-group. This 
new specimen possesses characters suggesting that the Honduran population is phylogenetically 
distinct from other C. goodwini, and we herein describe a new subspecies for C. goodwini. 

Our morphological analysis indicates that enlargement of the forefeet and foreclaws is a 
trend found throughout the C. mexicana-group, reaching its xenith in members of the C. 
goldmani-group. This trend is accompanied by functional modifications of the forelimb skel- 
eton. Our phylogenetic analysis, based on 29 transition series, indicates that the C. mexicana- 
group forms a well-supported clade and justifies nesting of the C. goldmani-group within the 
C. mexicana-group. Topology of each of the shortest trees shows that the species previously 
recognized as subspecies of C. mexicana are paraphyletic with respect to the C. goldmani- 
group. 



Resumen 

Las musarafias de orejas pequenas del grupo de especies Cryptotis goldmani ocurren en 
habitats de altas elevaciones (> 1000 msnm) en Mexico y el norte de Centroamerica. Anter- 
iormente, se consideraba que este grupo se constaba de solamente dos especies (C. goldmani 
y Cryptotis goodwini) que se caracterizaban por la agrandacion excesiva de las patas y las unas 
delanteras. Filogeneticamente C. goldmani y C. goodwini se encontraban en el grupo de es- 
pecies Cryptotis mexicana sensu Choate (1970), un grupo que ademas incluia solomente C. 
mexicana con cuatro subespecies. Nuestra revision de estas musaranas indica que las subes- 
pecies de C. mexicana son taxones bien diferenciados que reconocemos como especies distintas 
(C. mexicana, Cryptotis nelsoni, Cryptotis obscura, y Cryptotis peregrina). El grupo de especies 

IELDIANA: ZOOLOGY, N.S., NO. 91, DECEMBER 31, 1998, PP. 1-35 



C. goldmani consta de por lo menos cuatro especies distintas: Cryptotis alticola, en Colima, 
Jalisco, Mexico, Michoacan, Puebla y el Distrito Federal de Mexico; C. goldmani, en la Sierra 
Madre del Sur en Guerrero y Oaxaca; C. goodwini, en Chiapas, Guatemala, El Salvador, y 
Honduras; y Cryptotis griseoventris, en las montanas del norte de Chiapas y Guatemala. Un 
ejemplar unico de C. goodwini muestra la presencia de tal en Honduras, fuera de la distribution 
anteriormente conocida para el grupo de especies C. mexicana. Este especimen nuevo tiene 
caracteristicas que sugieren que la poblacion hondurena es filogeneticamente distinta a los 
demas C goodwini; por lo tanto, se lo describe aqui como una nueva subespecie de C. good- 
wini. ., 

Nuestro analisis morfologico muestra que la agrandacion de las patas y las unas delanteras 
es una tendencia presente en todo el grupo C. mexicana, y que alcanza su desarrollo maximo 
en las especies del grupo C. goldmani. Tal tendencia se relaciona con modificaciones funcion- 
ales del esqueleto del miembro delantero. Nuestro analisis filogenetico, basado en 29 caracteres, 
indica que el grupo C. mexicana es un clado bien definido y justifica la position del grupo C. 
goldmani dentro del grupo anterior. La topologia de cada uno de los arboles filogeneticos mas 
cortos demuestra que las especies anteriormente tratadas como subespecies de C. mexicana son 
parafileticas con respecto al grupo de especies C. goldmani. 



Introduction 

Small-eared shrews of the New World genus 
Cryptotis range geographically from southernmost 
Canada, through the eastern half of the United 
States, Mexico, and Central America, to the An- 
dean highlands of northwestern South America. A 
large array of names has been proposed for the 
Mexican and Central American members of the 
genus; however, the paucity of specimens avail- 
able has hindered assessment of geographic and 
nongeographic variation. Choate (1970), in his 
monograph on the Middle American members of 
this group, recognized eight species that he ar- 
ranged into three informal groupings: the "Cryp- 
totis mexicana- group" (containing three species: 
Cryptotis goldmani, Cryptotis goodwini, and C. 
mexicana), the "Cryptotis parva-group" (two 
species: C. parva and Cryptotis nigrescens), and 
"relict species" (three species: Cryptotis endersi, 
Cryptotis gracilis, and Cryptotis magna). Recent- 
ly, we (Woodman & Timm, 1992) described a 
ninth species for the region, Cryptotis honduren- 
sis, from high-elevation pine forest and pine-oak 
forest of south-central Honduras. In addition, we 
(Woodman & Timm, 1993) recognized C. nigres- 
cens, a member of Choate's C. /?arva-group, to be 
a complex of at least five species distributed from 
southern Mexico to Colombia. Thus, the genus 
Cryptotis is considerably more diverse in Central 
America than previously thought. 

The three species recognized by Choate (1970) 
as comprising his C. mexicana- group (C. gold- 
mani, C. goodwini, and C. mexicana) occur in 



middle- to high-elevation habitats distributed from; 
the Mexican state of Tamaulipas to El Salvador. 
Among these species, C. goldmani and C. good- 
wini can be distinguished most readily from C 
mexicana by their greatly enlarged front feet and 
foreclaws, characters unique among members of 
the genus. Under Choate's (1970) taxonomy, C. 
goldmani included two subspecies, C. goldmani^ 
alticola and C. goldmani goldmani, that werd 
known to occur from the Mexican state of Jalisco 
southeast into Guatemala. The monotypic C\ 
goodwini was known from Guatemala, El Salva- 
dor, and Chiapas, Mexico. 

We recently obtained a new specimen of the C. 
goldmani-group from Honduras, which is outside 
of the previously known geographic range for 
these broad-clawed shrews. Our attempts to iden- 
tify this specimen and adequately document its 
phylogenetic and biogeographic relationships to 
other members of the C. goldmani-group led us 
to reevaluate the taxonomic status of these poorly 
known shrews. 

It is the purpose of this paper to review pub- 
lished knowledge of the broad-clawed shrews of 
the C. goldmani-group and analyze their phylo- 
genetic relationships. We redescribe four species 
described previously (Cryptotis alticola, C. gold- 
mani, C. goodwini, and Cryptotis griseoventris) 
and describe a new subspecies for C. goodwini., 
Our accounts for these species include a review 
of all available information on reproduction, ele- 
vational and geographic distributions, and habitat, 
and we provide a key for their identification. We 
present diagnostic characteristics of the postcra- 
nial skeleton of Cryptotis for the first time. 






FIELDIANA: ZOOLOGY 




Fig. 1 . Left humerus of Cryptotis goldinani, indicating 
natomical features mentioned in the text. Abbrevia- 
pns: CA, capitulum: DP. deltoid process; GT. greater 
(jberosity; HD, head; LE, lateral epicondyle; ME, me- 
lial epicondyle; PP, pectoral process; TR, trochlea; and 
JT. teres tubercle. Anatomical terminology follows Reed 
11951). 



vlethods 

1 For the purposes of this paper, we use Cryptotis 
itexicana-group to refer informally to the set of 
axa that includes C. mexicana, Cryptotis nelsoni, 
tryptotis obscura, Cryptotis peregrina, Cryptotis 
llticola, Cryptotis goldmani, Cryptotis goodwini, 
nd Cryptotis griseoventris. The Cryptotis gold- 
jjiam'-group, or broad-clawed shrews, refers to the 
jjubset of the C. mexicana-group that includes C. 
llticola, C. goldmani, C. goodwini, and C. gri- 
seoventris. The Cryptotis nigrescens-group in- 
cludes Cryptotis colombiana, Cryptotis honduren- 
tis, Cryptotis mayensis, Cryptotis mera, Cryptotis 
\ierriami, and Cryptotis nigrescens (Woodman & 
j'imm, 1993). In the Cryptotis parva-group, we 
liclude all taxa currently included as subspecies 
f C. parva (C. parva berlandieri, C p. elasson, 
l. p. floridana, C. p. orophila, C. p. parva, C. p. 
aieblensis, soricina, and tropicalis). The C. mex- 
:ana-group and the C. goldmani-group appear to 
e natural groupings of species. The C. nigres- 
ens-group and the C. parva-group may represent 
lonophyletic groupings of species, but they re- 
lain informal, awaiting a comprehensive phylo- 
enetic study of the genus. 

Terminology of dentition and dental character- 
ises follows Choate (1970). Anatomical termi- 
ology of the humerus (Fig. 1) and other parts of 
le postcranial skeleton follows Reed (1951). 
leasurements used in our analyses follow Wood- 
lan and Timm (1993). Measurements of the skull 
ig. 2) were taken to the nearest 0.1 mm using 



either an ocular micrometer in a binocular micro- 
scope or a handheld dial caliper. Standard external 
measurements are those recorded by the collector, 
except for head-and-body length (HB), which we 
calculated by subtracting the recorded tail length 
(TL) from the total length. Other abbreviations of 
measurements found in the text include distance 
from articular condyle to posterior edge of M, 
(AC3); breadth of articular condyle (BAC); cra- 
nial breadth (CB); condylobasal length, not in- 
cluding the upper incisors (CBL); height of artic- 
ular condyle (HAC); height of coronoid process 
(HCP); height of coronoid valley (HCV); 
interorbital breadth (IO); length of lower first mo- 
lar (M,L); breadth of palate across second molars 
(M 2 B); length of molariform toothrow, P 4 through 
M 3 (MTR); mandibular length, from inferior sig- 
moid notch to mental foramen (ML); palatal 
length (PL); upper toothrow length, U 1 through 
M 3 (TR); lower toothrow length, P 3 through M, 
(TRD); breadth of palate across first unicuspids 
(U'B); breadth of palate across third unicuspids 
(U 3 B); length of unicuspid toothrow (UTR); and 
breadth of zygomatic plate (ZP). All measure- 
ments are in millimeters. Univariate statistics in- 
clude mean ± standard deviation. Multivariate 
analyses and analyses of variance (ANOVA) were 
carried out using BMDP on the University of 
Kansas Academic Computing Service IBM 
VM/CMS computer system and Minitab release 
8.0 for DOS-based personal computers. F-values 
and P-values for ANOVAs (Table 1) were cal- 
culated using a Brown-Forsythe equality of 
means test that does not assume equality of the 
variances. Localities and elevations that we report 
are taken directly from specimen tags and repre- 
sent the descriptions provided by the original col- 
lectors. We provide corrections and additions in 
brackets. Latitudinal and longitudinal coordinates 
for specimens used in the analysis of geographic 
variation of C. goodwini and C. griseoventris are 
from the U.S. Board on Geographic Names 
(1956a,b, 1965). Species and subspecies synony- 
mies list only published uses of names. 

We attempted to accurately characterize varia- 
tion in pelage coloration in our descriptions of 
species; all capitalized color names follow Ridg- 
way (1912). However, foxing is likely to have oc- 
curred, expecially in older specimens, and minor 
variations in pelage color should not be miscon- 
strued as real differences among species without 
further testing. In general, pelage in living mem- 
bers of the C. mexicana-group is dark chocolate- 
brown or darker in coloration, and it is not pos- 



/OODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 




Fig. 2. Measurements of the skull used in this study (see Table 1). The lines represent the orientation of the 
measurements. Abbreviations: AC3, distance from articular condyle to posterior edge of M 3 ; BAC, breadth of articular 
condyle; CB, cranial breadth; CBL, condylobasal length, not including the upper incisors; HAC, height of articular 
condyle; HCP, height of coronoid process; HCV, height of coronoid valley; IO, interorbital breadth; mlL, length 
of lower first molar; M2B, breadth of palate across second molars; MTR, length of molariform toothrow, P 4 through 
M 3 ; ML, mandibular length, from inferior sigmoid notch to mental foramen; PL, palatal length; TR, upper toothrow 
length, U 1 through M 3 ; TRD, lower toothrow length, P, through M,; U1B, breadth of palate across first unicuspids; 
U3B, breadth of palate across third unicuspids; UTR, length of unicuspid toothrow; ZP, breadth of zygomatic plate. 
Measurements follow those of Woodman and Timm (1993). 



sible to distinguish them on the basis of pelage 
alone. 

Previous investigations of sexual dimorphism 
in Cryptotis (Choate, 1970 — C. mexicana; Wood- 
man, 1992 — Cryptotis gracilis; Woodman & 
Timm, 1993 — C nigrescens) revealed no clear 
pattern of morphological differences between 
males and females. We had insufficient sample 
sizes of complete individuals of both sexes to car- 
ry out adequate analyses of secondary sexual vari- 
ation for any one species. Based on the lack of 
obvious sexual variation in these or other mem- 
bers of the genus, we assumed that there was no 
confounding sexual variation in our analyses. 
Therefore, males, females, and specimens of un- 
known sex were combined. 

In our investigations of the C. goldmani-group, 
we operated under the philosophical influence of 
the evolutionary species concept as redefined by 
Wiley (1978, 1981). This concept requires that the 
species category (and taxonomy in general) be 
logically consistent with reconstructed phyloge- 
netic history. In application, we considered a spe- 
cies as the largest monophyletic entity whose con- 
stituent parts interact and are not on different evo- 



lutionary trajectories (Alternative 2 of Frost & 
Hillis, 1990). Operationally, we used unique dis- 
tributional patterns of morphological characters 
among populations to distinguish presumably ge- 
netically cohesive groups from other genetically 
cohesive groups. This results in a clear, support- 
able, testable framework for further phylogenetic, 
biogeographic, and ecological studies. Similarly, 
we consider the subspecies category to represent 
a phylogenetically distinct entity; in practice, we 
use it to separate poorly defined taxa that require 
further taxonomic study. 

Phylogeny and character evolution in the C. 
goldmani-group were analyzed using PAUP 3.1.1 
(see Swofford, 1993) and MacClade 3.0 (see 
Maddison & Maddison, 1992). Phylogenetic anal- 
ysis was carried out using an exhaustive search of 
29 unordered transition series (TS; see Appendix 
I). In addition to the four species in the C. gold- 
mani-group, we included the four taxa previously 
treated as subspecies of Cryptotis mexicana (C. 
mexicana, C. nelsoni, C obscura, and C pere- 
grind). One or more of these taxa should represent 
the first outgroup to the C goldmani-group, and 
we thought that the C goldmani-group might 



FIELDIANA: ZOOLOGY 



>rove to be paraphyletic with respect to one or 
nore of these taxa. As outgroups, we chose C. 
yarva parva from Kansas and C. nigrescens from 
Dosta Rica because they were considered previ- 
>usly to be phylogenetically distinct from the C. 
nexicana-group (Choate, 1970), and large series 
)f specimens of those taxa were readily available 
o us. 

Specimens from the following institutions were 
lsed in this study: American Museum of Natural 
iistory, New York (amnh); Angelo State Univer- 
ity Natural History Collection, San Angelo 
asnhc); California Academy of Sciences, San 
-rancisco (cas); Escuela Nacional de Ciencias 

iologicas, Mexico City (encb); Field Museum, 
Chicago (fmnh); Instituto de Biologia, Universi- 
lad Nacional Autonoma de Mexico, Mexico City 
ibunam); Instituto Nacional de Biodiversidad, 
Janto Domingo de Heredia, Costa Rica (iNBio); 
Jniversity of Kansas Natural History Museum, 
^awrence (ku); Natural History Museum of Los 
Vngeles County, Los Angeles (lacm); Museum of 
Comparative Zoology, Cambridge (mcz); James 

brd Bell Museum of Natural History, St. Paul 
mmnh); Museo Nacional de Costa Rica, San Jose 
mncr); Museo de Zoologia, Facultad de Ciencias, 
Jniversidad Nacional Autonoma de Mexico, 
Mexico City (mzfc); Forschungsinstitut und Na- 
urmuseum Senckenberg, Frankfurt am Main 
smf); Texas Cooperative Wildlife Collection, 
College Station (tcwc); University of Michigan 
Museum of Zoology, Ann Arbor (ummz); and Na- 
ional Museum of Natural History, Washington 
usnm). 



Systematic Biology 



Based on elevational and geographic distribu- 
ions of specimens and the correlation of character 
tates to these elevational and geographic distri- 
•utions, we recognize four distinct species within 
he C. goldmani-group. These are C. goodwini, C. 
".oldmani, and two species previously synony- 
nized with C. goldmani (C. alticola and C. gri- 
eoventris). In addition, we recognize C. goodwini 
-s a polymorphic species consisting of C. g. good- 
vini and a new subspecies that we describe below. 



Family Soricidae von Waldheim, 1817 
Subfamily Soricinae von Waldheim, 1817 
Genus Cryptotis Pomel, 1848 

"Cryptotis mexicana -group" Choate, 1970 

Description — These shrews are small to me- 
dium-sized members of the genus with long, dark 
dorsal pelage. They are characterized by variably 
broadened forefeet; moderately elongate and 
moderately broad to greatly elongate and very 
broad foreclaws; posterior border of the zygo- 
matic plate usually positioned equal to, or slightly 
posterior to, the posterior base of the maxillary 
process (Fig. 3C); upper toothrow uncrowded; 
dentition not bulbous; anterior border of the cor- 
onoid process of the mandible joins the horizontal 
ramus at a relatively low angle; posterior border 
of lower incisor extends to posterior cingulum of 
P 4 ; relatively long distance from the coronoid pro- 
cess to the posterior border of M 3 ; tall, wide ar- 
ticular face of the articular process; deep lower 
sigmoid notch; relatively long, low P,; relatively 
short, broad metacarpals; shortened and broad- 
ened humerus with enlongated processes and a 
dorsoventrally elongate head; and deeply pocket- 
ed posterior edge of the falciform process of the 
tibia. 

Included Taxa — Cryptotis alticola, C. gold- 
mani, C. goodwini, C. griseoventris, C. mexicana, 
C. nelsoni, C. obscura, and C. peregrina. 



"Cryptotis goldmani-group" 

Description — A subset of the C. mexicana- 
group, the C. goldmani-group consists of medi- 
um-sized members of the genus with relatively 
short tails (mean TL < 39% of HB— Tables 1 and 
2); greatly broadened forefeet; extremely long, 
broad foreclaws; fourth upper unicuspid usually 
aligned with the unicuspid toothrow and partially 
visible in labial view; protoconal basin of M 1 re- 
duced relative to hypoconal basin; M 1 simple, hy- 
pocone absent or poorly developed and lacking 
metacone; relatively low coronoid process of the 
mandible; entoconid of M, vestigial or absent; and 
extremely broad humerus, with greatly enlongated 
processes. 

Included Taxa — Cryptotis alticola, C. gold- 
mani, C. goodwini, and C. griseoventris. 



VOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



Table 1. Measurements of Cryptotis used in this study. Statistics presented are mean ± standard deviation of 
the mean and observed extremes. Sample sizes (n) are different for skin measurements, skull measurements, and 
weight; sample sizes for cranial breadth are different than those for other variables of the skull. The F-value and P- 
value for each variable are from an analysis of variance (ANOVA) of the variables from C. goldmani and the two 
species previously considered conspecific with it (C. alticola and C. griseoventris). 









Cg. 
goodwini 


Cg. 




griseoventris 


C. alticola 


C. goldmani 


C. g. goodwini 


(El 
Salvador) 


magnimana 
(holotype) 


C. 


External measurements 














n = 22 


n = 28 


n = 34 


n = 1 


n = 1 




n = 27 


Head-and-body length 

79 ± 5 


76 ± 5 


84 ± 5 




80 




77 ± 3 


69-87 


62-84 


k 75-94 








73-85 


Tail length 

26 ± 2 


29 ± 3 


29 ± 2 




25 




29 ± 1 


23-30 


24-36 


25-34 








27-32 



Skull measurements 

n = 16 n = 16 n = 20 
Condylobasal length (F = 6.53, P = 0.003) 

20.2 ± 0.5 19.6 ±0.5 21.1 ± 0.7 

19.3-21.1 18.9-20.5 20.0-22.8 
Cranial breadth 

10.4 ± 0.2 10.2 ±0.2 11.1 ± 0.3 

9.9-10.8 9.8-10.5 10.6-11.6 

(n = 15) (n = 18) (n = 15) 
Breadth of zygomatic plate (F = 1.91, P = 0.164) 

1.9 ± 0.1 1.8 ± 0.2 1.9 ± 0.1 

1.7-2.1 1.4-2.2 1.6-2.2 
Interorbital breadth (F = 4.92, P = 0.012) 

4.9 ± 0.2 5.0 ± 0.2 5.6 ± 0.2 

4.7-5.3 4.7-5.3 5.3-5.8 
Breadth across first unicuspids (F = 31.53, P = 0.0000) 

2.7 ± 0.1 2.6 ± 0.1 2.7 ± 0.1 

2.5-2.8 2.4-2.7 2.6-2.9 
Breadth across third unicuspids (F = 19.51, P = 0.0000) 

3.2 ± 0.1 3.0 ± 0.1 3.3 ± 0.1 

3.0-3.2 2.8-3.2 3.0-3.5 
Breadth across second molars (F = 95.12, P = 0.0000) 

6.2 ± 0.2 5.8 ± 0.1 6.2 ± 0.2 

5.8-6.4 5.6-5.8 6.0-6.6 
Palatal length (F = 0.20, P = 0.816) 

8.7 ± 0.3 8.7 ± 0.3 9.2 ± 0.3 

8.1-9.2 8.0-9.2 8.8-10.1 
Length of upper toothrow (F = 4. 72, P = 0.014) 

7.6 ± 0.2 7.4 ± 0.2 7.9 ± 0.3 

7.1-7.9 7.0-7.8 7.5-8.6 
Length of unicuspid toothrow (F = 55.19, P = 0.0000) 

2.5 ± 0.1 2.3 ± 0.1 2.7 ± 0.1 

2.3-2.7 2.0-2.5 2.5-2.9 
Length of molariform toothrow (F = 17.71, P = 0.0000) 

5.5 ± 0.2 5.4 ± 0.1 5.6 ± 0.2 
5.1-5.7 5.3-5.7 5.3-5.9 

Length of mandible (F = 14.35, P = 0.0000) 

6.6 ± 0.2 6.3 ± 0.3 6.6 ± 0.2 
5.9-6.8 5.8-6.8 6.2-6.8 

Height of coronoid process (F = 16.35, P = 0.0000) 

4.6 ± 0.2 4.6 ± 0.2 4.8 ± 0.1 

4.3-4.9 4.3-4.9 4.6-5.2 



20.8 



10.8 



5.6 



2.8 



3.3 



6.2 



9.1 



7.8 



2.5 



5.7 



4.8 



20.3 



10.8 



2.2 



5.3 



2.6 



3.0 



5.9 



8.8 



7.5 



2.4 



5.5 



6.5 



4.7 



n = 20 



19.9 ± 0.4 


18.8- 


-20.4 


10.2 ± 0.2 


9.8- 


-10.7 


(n = 


19) 


1.9 ± 0.1 


1.6-2.1 


5.1 ± 0.2 


4.8- 


-5.3 


2.5 ± 0.1 


2.3- 


-2.6 


3.0 ± 0.1 


2.8- 


-3.2 


5.6 ± 0.1 


5.4- 


-5.9 


8.7 ± 0.2 


8.1- 


-9.0 


7.6 ± 0.2 


7.3- 


-7.8 


2.7 ± 0.1 


2.5- 


-2.9 


5.2 ± 0.1 


5.0- 


-5.4 


6.2 ± 0.2 


5.7- 


-6.4 


4.4 ± 0.1 


4.2- 


-4.6 



FIELDIANA: ZOOLOGY 



Table 1. Continued. 



C. alticola 


C. goldmani C 


g. goodwini 


e.g. 

goodwini 

(El 
Salvador) 


e.g. 

magnimana 
(holotype) 


C. griseoventris 


Height of coronoid valley {F - 
2.9 ± 0.1 
2.8-3.1 


10.75. P = 0.0001) 

2.8 ± 0.1 

2.7-3.0 


3.0 ± 0.1 
2.8-3.4 


2.8 


2.9 


2.8 ± 0.1 
2.6-3.0 


Height of articular condyle (F = 
4.1 ± 0.2 
3.7-4.3 


= 9.22, P = 0.0005) 

3.9 ± 0.2 

3.8-4.2 


4.2 ± 0.2 
3.8-4.6 


4.1 


4.1 


3.9 ± 0.1 
3.7-4.1 


Breadth of articular condyle (F 
3.3 ± 0.1 
3.0-3.5 


= 31.46. P = 0.0000) 
3.1 ± 0.1 
2.9-3.2 


3.3 ± 0.2 
3.0-3.6 


3.3 


3.1 


3.0 ± 0.1 
2.8-3. 1 


Articular condyle to M, (F = 3.64, P = 0.035) 
5.2 ± 0.2 5.1 ± 0.2 
4.8-5.6 4.8-5.3 


5.6 ± 0.2 
5.2-5.9 


5.6 


5.3 


5.1 ± 0.1 
4.8-5.3 


Length of lower toothrow (F = 
6.2 ± 0.2 
5.8-6.5 


1.11, P = 0.341) 
6.1 ± 0.2 
5.4-6.4 


6.4 ± 0.2 
6.1-6.8 


6.5 


5.9 


6.1 ± 0.1 
5.8-6.3 


Length of M, 

1.8 ± 0.1 
1.7-2.0 


1.8 ± 0.1 
1.7-2.0 


1.9 ± 0.1 
1.8-2.0 


2.0 


1.8 


1.8 ± 0.1 
1.7-1.9 


Weight (g): 

11 ± 3 
8-16 
(n = 10) 


8 ± 1 

6-10 

(n = 15) 


16 ± 2 

12-19 

(n = 10) 


— 


— 


— 




I 1 

1 mm 



zygomatic 
plate 





Fig. 3. Lateral views of the left zygomatic plates of A, Cryptotis parva orophila; B, C. nigrescens; and C, C. 
mexicana. The arrangement of the zygomatic plate in C. mexicana, in which the posterior border is even with and 
confluent with the posterior root of the maxillary process, is representative of the C. mexicana-group. Among members 
of the C. nigrescens-group, the posterior border of the zygomatic plate and the posterior root of the maxillary process 
are separated by the posterior edge of the palate. Scale bar = 1 mm. 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



Table 2. Comparisons of characters among taxa in the Cryptotis goldmani-group. Relative measurements are 
percentages. Sample sizes are presented in accounts for each species. Abbreviations as in Methods section of the 
text. 



C. alticola 


C. goldmani 


C. g. goodwini 


C. g. magnimana 


C. griseoventris 


Foramen of sinus canal 










vestigial 


well developed 


absent 


absent 


absent 


Foramen dorsal to dorsal articular facet present 








39% 


16% 


90% 


present 


81% 


Two dorsal foramina present 










54% 


61% 


82% 


present 


92% 


Posteroventral border of unicuspids 










straight 


concave 


concave 


concave 


concave 


to convex 










Vestigial entoconid of M, present 










64% 


52% 


14% 


? 


0% 


Humerus 










derived 


most derived 


derived 


derived 


? 


(Fig. 15E) 


(Fig. 15H) 


(Fig. 15G) 


(Fig. 15F) 




Tail length as % of head-and-body length (TL/HB X 100) 








33 ± 4 


38 ± 5 


35 ± 3 


31 


38 ± 2 


27-41 


31-48 


30-41 




34-42 


Relative length of rostrum (PL/CBL 


X 100) 








43.2 ± 1.2 


44.2 ± 1.0 


43.6 ± 0.7 


43.3 


43.8 ± 0.8 


40.1-45.5 


41.9-45.6 


42.5-44.8 




42.1-45.0 


Relative breadth of zygomatic plate 


(ZP/PL X 100) 








21.8 ± 1.2 


21.1 ± 2.4 


21.1 ± 1.9 


25.0 


22.3 ± 1.7 


19.6-24.1 


15.7-23.4 


17.2-24.4 




18.2-24.7 


Relative length of unicuspid toothrow (UTR/CBL X 100) 








12.3 ± 0.4 


11.9 ± 0.6 


13.0 ± 0.5 


11.8 


13.7 ± 0.5 


11.6-13.0 


10.5-13.0 


12.0-14.0 




12.9-14.5 


Relative palatal breadth (M2B/PL X 


100) 








70.9 ± 3.0 


66.4 ± 2.3 


67.6 ± 2.4 


67.1 


64.3 ± 2.1 


64.1-77.8 


63.0-71.2 


62.4-71.6 




60.7-68.2 


Relative height of coronoid process 


(HCP/ML X 100) 








70.5 ± 3.3 


73.6 ± 4.4 


72.7 ± 2.2 


72.3 


70.5 ± 1.8 


64.2-80.0 


66.2-81.0 


68.7-76.5 




67.2-73.7 


Relative length of posterior portion of the mandible (AC3/ML X 100) 






79.3 ± 2.0 


80.5 ± 2.6 


84.2 ± 2.8 


81.5 


82.9 ± 2.3 


76.2-83.6 


76.1-84.5 


79.1-90.5 




79.7-86.9 



Cryptotis goodwini Jackson, 1933 
Cryptotis goodwini goodwini Jackson, 1933 

Cryptotis goodwini Jackson, 1933:81; Goodwin, 
1934:6; Felten, 1958:218; Hall and Kelson, 
1959:61; Genoways and Choate, 1967:204; 
Choate, 1970:249; Hall, 1981:60; Medellin, 
1988:84. 

Cryptotis nigrescens: Burt and Stirton, 1961:21 
(in part). 

C[ryptotis]. goodwini: Musser, 1964:7. 



Holotype — Skin and skull of adult male, usnm 
77074; taken 13 January 1896 by E. W. Nelson 



(collector number 9073); from "Calel" (Jackson 
1933:81), Quezaltenango, Guatemala, 10,200 ft. 

Distribution — Contiguous highland regions 
above 1100 m in the Mexican state of Chiapas 
and Guatemala, and in a separate highland area in 
northern El Salvador. The population in El Sal- 
vador probably extends northward along the Mer- 
endon Cordillera into western Honduras (Fig. 4). 

Description — Size large for the genus (Table 
1); tail short, averaging 29 mm, or ca. 35% (Table, 
2) of HB; the dorsal guard hairs 6-8 mm long; 
dorsal pelage Saccardo's Umber/Prout's Brown/ 
Mummy Brown; ventral pelage somewhat paler 
than dorsum, Light Drab/Drab to almost Wood 
Brown; dorsal fur three-banded: basal five-sixths 



FIELDIANA: ZOOLOGY 




WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



of hairs silvery gray, followed by a thin, very pale 
brown band grading into dark brown tip. Rostrum 
of moderate length (PL/CBL = 43.6% ± 0.7, n 
= 20); usually two well-developed dorsal foram- 
ina (82%, n = 28); no ventral extension of the 
sinus canal or associated foramen posterior to dor- 
sal articular facet (see account for C. goldmani, 
below); a foramen dorsal to the dorsal articular 
facet usually present on one or both sides of the 
skull (90%, n = 20); zygomatic plate of moderate 
breadth (ZP/PL = 21.1% ± 1.9, n = 20), anterior 
border usually aligned with mesostyle-metastyle 
valley or metastyle of M 1 ; P 4 , M 1 , and M 2 slightly 
to moderately recessed on posterior border; M 3 
usually with paracrista, paracone, precentrocrista, 
mesostyle, very short postcentrocrista (all pig- 
mented), and a well-developed, pigmented proto- 
cone. Mandible relatively long and of moderate 
breadth for the genus; articular process generally 
moderately tall and wide, with a moderately broad 
lower articular facet; entoconid usually absent 
(86%, n = 29), but when present, poorly devel- 
oped. 

Comparisons — Cryptotis goodwini goodwini is 
the largest of the broad-clawed shrews in overall 
body size. 

Cryptotis alticola — Cryptotis g. goodwini has 
somewhat shorter, narrower foreclaws; lacks ven- 
tral extension of sinus canal and associated fora- 
men posterior to dorsal articular facet; has a fo- 
ramen dorsal to dorsal articular facet; has a rela- 
tively narrower palate; and has a shorter mandi- 
ble. 

Cryptotis goldmani — Cryptotis g. goodwini has 
a relatively shorter tail; lacks ventral extension of 
sinus canal and associated foramen posterior to 
dorsal articular facet; has a foramen developed 
dorsal to the dorsal articular facet; usually lacks 
an entoconid on M 3 ; and has a longer, narrower 
humerus with less enlarged bony processes. 

Cryptotis griseoventris — Cryptotis g. goodwini 
is larger overall. 

Remarks — The known biology of Cryptotis 
goodwini goodwini was summarized previously 
by Choate (1970) and Choate and Fleharty (1974). 
Cryptotis goodwini goodwini has been found in 
high-elevation pine forests, sometimes mixed with 
firs, oaks, and other trees. Severe frosts are as- 
sociated with at least some of their habitats (Gold- 
man, 1951). Most specimens are from localities 
of 1200 m or more in elevation. Although Choate 
(1970:251) reported a specimen from Finca Xic- 
acao, Guatemala, as from ca. 3000 ft, no elevation 
is given on the specimen tag for this specimen, 



and Xicacao is located in a region higher than 
1000 m. James W. Bee (ku unpublished field cat- 
alog, 1954, 1955) recorded taking Sorex verae- 
pacis, Microtus guatemalensis, Peromyscus gua- 
temalensis, Reithrodontomys microdon, Reithro- 
dontomys sumichrasti, and Reithrodontomys ten- 
uirostris along with C. g. goodwini 3.5 mi SW of 
San Juan Ixcoy, Guatemala, at 10,120 ft, on 27 
December 1954. At 6000 ft, 5 mi N and 1 mi W 
of El Choi, Guatemala, on 30 January 1955, he 
captured C. g. goodwini with Oligoryzomys ful- 
vescens, Peromyscus aztecus, Peromyscus levipes, 
Reithrodontomys fulvescens, Reithrodontomys 
mexicanus, R. sumichrasti, and Scotinomy tegui- 
na. 

Reproductive data for Cryptotis goodwini 
goodwini are lacking. Our inspection of study 
skins of males taken from 4 to 15 January (n = 
6), 5-22 May (n = 2), 20 August (n = 3), and 27 
December (n = 1) yielded none with well-devel- 
oped lateral glands. On study skins, lateral glands 
typically appear as paired, oval regions approxi- 
mately 6-9 mm in length and 5-6 mm in width. 
These glandular areas lack long guard hairs and 
normal underfur, but they have a sparse covering 
of short, fine, pale hairs. Among most genera of 
Soricidae, both males and females possess lateral 
glands, but those of females are smaller and more 
difficult to see (Murariu, 1976; Bee et al., 1980). 
Eadie (1938) found that the lateral glands of Blar- 
ina showed increased activity with enlargement of 
the testes, and they may serve a function in sexual 
and social communication. The occurrence of en- 
larged lateral glands on study skins has been used 
to infer timing of sexual activity in adult male 
Cryptotis (Woodman & Timm, 1993). 

The occurrence of Cryptotis goodwini goodwini 
in El Salvador is documented by a single speci- 
men from Hacienda Montecristo on Cerro Mon- 
tecristo (Felten, 1958; Choate, 1970). Cerro Mon- 
tecristo is an isolated mountain, separated from 
the Guatemalan Highlands by a deep, wide valley 
occupied by tributaries of the Rio Motagua. This 
lowland valley supports xeric vegetation (Stuart, 
1954), unlike that found in cool, moist highlands 
generally inhabited by C. g. goodwini, and it 
probably provides an effective barrier to gene 
flow. Because of this, we hypothesized that the 
specimen from El Salvador might prove to rep- 
resent a distinct species. This individual has less 
emarginate upper dentition than C. g. goodwini, 
but otherwise it cannot be separated from them 
on qualitative characteristics. To test the overall 
similarity of the specimen from El Salvador to 



10 



FIELDIANA: ZOOLOGY 



those from Guatemala, we carried out a principal 
components analysis using seven uncorrelated 
variables (CBL, ZP, 10, M 2 B, UTR, MTR, and 
HCP) from 20 Guatemalan C g. goodwini, plus 
the specimen from El Salvador. A plot of the first 
three factor axes from this analysis is shown in 
Figure 5. The specimen from El Salvador shows 
up in the centers of factor axis 1 (size; see Table 
6) and factor axis 2 (UTR and HCP), but it rep- 
resents the lower extreme of variation for factor 
3 (ZP and UTR). This is due to the individual's 
relatively narrow zygomatic plate and its relative- 
ly short unicuspid toothrow (Table 1). The spec- 
imen from El Salvador falls within the overall 
range of variation for C. g. goodwini, and, lacking 
any definitive characters that would set it off from 
that species, we consider it most appropriate to 
continue to refer it to C. g. goodwini. 

Specimens Examined (36)— EL SALVADOR: 
SANTA ANA: Hacienda [Finca] Montecristo 
[13°40'N, 89°29'W] (smf 14837). GUATEMA- 
LA: ALTA VERAPAZ: Finca Xicacao (ummz 
87869). BAJA VERAPAZ: 5 mi N, 1 mi W El 
Choi, 6000 ft (KU 64611). CHIMALTENANGO: 
Santa Elena [14°48'N, 91°01'W], 9900-10,000 ft 
(fmnh 41791-41794). HUEHUETENANGO: 3.5 
mi SW San Juan Ixcoy [15°36'N, 91°27'W], 
10,120 ft (ku 64610); San Mateo Ixtatan, ca. 4 
km NW Santa Eulalia [15°45'N, 91°29'W], 
Yayquich, 2950 m (ummz 117843); Hacienda 
Chancol, 9500-1 1,000 ft (usnm 77069). JALAPA: 
Mataquescuintla [14°32'N, 90°11'W], 8400 ft 
(usnm 275681). QUEZALTENANGO: Calel 
[15°04'N, 91°34'W], 10,200 ft (usnm 77070, 
77072-77073, 77075-77084, includes holotype); 
Volcan Santa Maria [14°45'N, 91°33'W], 9000- 
11,000 ft (usnm 77086-77087). SAN MARCOS: 
S slope Volcan Tajumulco, 10,000 ft (ummz 
99541). TOTONICAPAN: Cumbre Maria Teciin 
[14°52'N, 91°13'W], 3000 m (ummz 112004- 
1 1201 1). MEXICO: CHIAPAS: 17 km SE Finca 
Prusia, Reserva Ecologica El Triunfo, 2000 m 
(ibunam 22784). 

Additional Records— GUATEMALA: CHI- 
MALTENANGO: Tecpam [= Tecpan], 9700 ft 
(Goodwin, 1934); SAN MARCOS: Finca La Paz, 
1200 m (Choate, 1970:251). 

Cryptotis goodwini magnimana, new subspecies 
Honduran Broad-clawed Shrew 
(Fig. 6) 

Holotype — Fluid-preserved pregnant adult fe- 
male with skull and humerus removed and 



cleaned, ku 144611; collected 21 November 1991 
by Peter Holm and Gustavo Cruz (no field num- 
ber). Fluid-preserved body intact; skull nearly 
complete, but with a hole in braincase dorsally 
and lacking left tympanic. 

Type Locality — 2.5 km N, 1.6 km E Cerro San 
Juanillo [14°30'N, 87°53'W], Reserva Biologica 
Cordillera de Montecillos, Comayagua Depart- 
ment, Honduras, 1730 m (see Fig. 4). The only 
known specimen was found dead in mixed pine 
and broadleaf forest. 

Distribution — Known only from the type lo- 
cality; probably occurs throughout the Cordillera 
de Montecillos and adjoining highland areas in 
high-elevation (> 1000 m) pine forest and mixed 
pine and broadleaf forest (Fig. 4). 

Measurements of Holotype — HB, 80; TL, 25; 
hindfoot, 14; ear, 6; CBL, 20.3; PL, 8.8; M 2 B, 5.9; 
IO, 5.3; TR, 7.5; CB, 10.8; TRD, 5.9 (see Table 
1). Because no external measurements were re- 
corded by the collectors, we measured the fluid- 
preserved body prior to removal of the skull. 

Diagnosis — A taxon possessing all of the com- 
mon characteristics of the Cryptotis goldmani- 
group (see above). Within this group, the holotype 
is characterized by its relatively short tail; rela- 
tively broad zygomatic plate; lack of a ventral ex- 
tension of the sinus canal; a well-developed fo- 
ramen dorsal to the dorsal articular facet; two 
well-developed dorsal foramina (= foramina or- 
bitalia of Biihler, 1964); relatively short upper 
unicuspid toothrow; unicuspid teeth with concave 
posteroventral border; P 4 , M 1 , and M 2 very slight- 
ly recessed on posterior border; and a short, broad 
humerus similar in grade to that of Cryptotis g. 
goodwini. 

Description — Size medium for the genus (Ta- 
ble 1); tail short, 25 mm (31% of HB— see Table 
2); dorsal guard hairs about 5-7 mm long; dorsal 
pelage of fluid-preserved holotype Mummy 
Brown to Fuscous when air-dried; Brownish Olive 
venter appears to be somewhat paler than dorsum; 
dorsal fur two-banded, basal three-fourths silvery 
gray, distal one-fourth grading from pale brown 
to brown at the tip. Rostrum of moderate length 
(PL/CBL = 43.3%); two well-developed dorsal 
foramina; no ventral extension of the sinus canal 
or associated foramen posterior to the dorsal ar- 
ticular facet (see account for C. goldmani, below); 
well-developed foramina dorsal to the dorsal ar- 
ticular facet on both sides of the skull; zygomatic 
plate broad (ZP/PL = 25%), anterior border 
aligned with metastyle of M 1 ; posterior border of 
P, M', and M 2 only very slightly recessed; M- 1 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



11 



Guatemala 
) El Salvador 



W 




Fig. 5. Three-dimensional plot showing similarities and differences among specimens of Cryptotis goodwini from 
Guatemala (solid squares) and El Salvador (open circle) based on the first three factors from principal components 
analysis of seven skull variables. Factor 1 represents size, factor 2 contrasts length of unicuspid toothrow with height 
of coronoid process, and factor 3 represents length of zygomatic plate and length of unicuspid toothrow (Table 6). 



has paracrista, paracone, precentrocrista, meso- 
style, very short postcentrocrista, and well-devel- 
oped protocone (because of wear, it cannot be de- 
termined whether protocone was pigmented). 
Mandible relatively long and of moderate breadth; 
articular process generally tall and wide, with a 
broad lower articular facet and a relatively shal- 
low lingual notch between articular facets; be- 
cause of wear, it cannot be determined whether an 
entoconid was present in talonid of M 3 . Humerus 
similar to C. g. goodwini (see below). 

Comparisons — Cryptotis goodwini magnimana 
has a relatively broader zygomatic plate and a rel- 
atively shorter unicuspid toothrow compared to 
those of the other four taxa of broad-clawed 
shrews. 

Cryptotis goodwini goodwini — Cryptotis g. 
magnimana is smaller overall (see Remarks, be- 
low); has a relatively and absolutely shorter uni- 
cuspid toothrow; and has less deeply recessed 
posterior borders of P 4 , M 1 , and M 2 . 

Cryptotis alticola — Cryptotis g. magnimana 
has somewhat shorter, narrower foreclaws; lacks 
ventral extension of sinus canal and associated fo- 
ramen posterior to dorsal articular facet; possesses 



a foramen dorsal to posterior dorsal articular facet 
and two well-developed dorsal foramina; has a 
relatively and absolutely narrower palate; has con- 
cave posteroventral borders of unicuspids; and has 
less deeply recessed posterior edges of P, M 1 , and 
M 2 . 

Cryptotis goldmani — Cryptotis g. magnimana 
is larger overall and has a relatively shorter tail; 
lacks ventral extension of sinus canal and asso- 
ciated foramen posterior to dorsal articular facet; 
possesses a foramen dorsal to dorsal articular fac- 
et and two well-developed dorsal foramina; and 
has a relatively longer, narrower humerus with 
less enlarged bony processes (see below). 

Cryptotis griseoventris — Cryptotis g. magni- 
mana is larger overall (see Remarks, below); has 
a relatively and absolutely shorter tail; has a rel- 
atively and absolutely broader zygomatic plate; 
and has a relatively and absolutely shorter uni- 
cuspid toothrow. 

Remarks — The only known specimen of this 
subspecies is an adult female that was pregnant 
when collected on 21 November 1991. She car- 
ried four embryos, three in the right horn of the 



12 



FIELDIANA: ZOOLOGY 




5 mm 



Fig. 6. Dorsal and ventral views of the cranium and lateral view of the skull of the holotype of Cryptotis goodwini 
magnimana. Scale bar = 1 mm. 



uterus and one in the left (crown to rump length 
= 8 mm). 

Cryptotis goodwini magnimana, C. goodwini 
goodwini, and C. griseoventris are similar in 
many respects, and the three taxa share several 
synapomorphies (see below) that distinguish them 
from C. alticola and C. goldmani. Many of the 
differences among C. g. magnimana, C. g. good- 
wini, and C. griseoventris are related to variation 
in body size or proportions of certain features (Ta- 
ble 2). To help evaluate the distinctiveness of 
these three populations, we examined their overall 
similarity using principal components analysis on 
eight variables (CBL, ZP, 10, UTR, MTR, HCP, 
BAC, and M,L) measured on 20 C. griseoventris, 
22 C. g. goodwini, and the holotype of C. g. mag- 
nimana. The specimens of C. griseoventris in- 
cluded seven from Chiapas and 13 from Guate- 
mala; those of C. g. goodwini included 21 from 
Guatemala and one from El Salvador. Each of 
these subsets was plotted separately to facilitate 
recognition of geographic variation within and 
among taxa. A plot of the first two factor axes 
from the principal components analysis (Fig. 7) 



shows all three named taxa to be distinct. Six of 
the eight variables are weighted relatively evenly 
on the first factor axis (Table 3), suggesting that 
this axis represents overall size of the individuals. 
Breadth of zygomatic plate (ZP) and length of un- 
icuspid toothrow (UTR) are weakly weighted on 
the first axis, indicating that they do not correlate 
strongly with size. Correlation coefficients of ZP 
and UTR with each of the other variables in the 
analysis fall below 0.24, and many are negative 
(Table 4). The first factor axis clearly separates 
the larger C. g. goodwini from the smaller C. gri- 
seoventris, with the holotype of C. g. magnimana 
between them. The difference in size between C. 
g. goodwini and C. griseoventris indicated by the 
first factor axis is supported by Mests between 
means of each of the individual variables; of 21 
variables tested, only three — length of tail (TL), 
breadth of zygomatic plate (ZP), and length of 
unicuspid toothrow (UTR) — did not show statis- 
tical differences between means for the two taxa. 
Means of each of the other 1 8 variables were sig- 
nificantly different (P < 0.001). On the second 
factor axis, ZP and UTR weigh out heavily, in- 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



13 



2 - 



1 - 



CM 



O 



-2 



-3 - 



O 



O 



o o 
o 
o 



<9 



o 



V 



▼ v T V 



— I — 
-4 



— I — 
-3 



-2 



-1 



factor 1 



C.g. goodwini: 

O Guatemala 

• El Salvador 

C. g. magniTnaTia: 

a Honduras 



C. griseoventris: 

V Chiapas 
T Guatemala 



Fig. 7. Plot of the first two factors from principal components analysis of specimens of Cryptotis griseoventris, C. 
goodwini goodwini, and Cryptotis g. magnimana. Factor 1 represents size; factor 2 contrasts length of zygomatic 
plate and length of unicuspid toothrow (Table 3). 



dicating that the axis represents a contrast be- 
tween these two variables (Table 3). The second 
factor axis separates the holotype of C. g. mag- 
nimana from C. g. goodwini and C. griseoventris, 
reflecting the relatively shorter unicuspid tooth- 
row and relatively broader zygomatic plate of C. 
g. magnimana. 

We subsequently investigated size variation 



Table 3. Factor loadings for the first two factor axes 
from principal components analysis of Cryptotis gri- 
seoventris, C. goodwini goodwini, and C. goodwini mag- 
nimana. These two axes (PCI, PC2) together accounted 
for 74% of the variation. Variables are listed in descend- 
ing order by their loadings on the first factor axis. Ab- 
breviations as in the Methods section. 



Variable 


PCI 


PC2 


ZP 


0.052 


-0.666 


UTR 


-0.048 


0.638 


mlL 


-0.349 


-0.263 


IO 


-0.408 


-0.075 


MTR 


-0.410 


0.145 


BAC 


-0.419 


0.056 


HCP 


-0.424 


-0.185 


CBL 


-0.427 


0.123 



among Cryptotis goodwini magnimana, C. g. 
goodwini, and C. griseoventris in relation to their 
geographic distributions to determine whether dif- 
ferences among them might be attributable to 
clinal variation. Cryptotis g. goodwini and C. gri- 
seoventris appear to be parapatric; their distribu- 
tions may overlap on a regional scale, but the two 
taxa have never been reported from the same site 
(Fig. 4). The distance between the closest known 
occurrences of C. g. goodwini and C. griseoven- 
tris is much less than the dimensions of either of 
their geographic ranges. If the two populations 
were members of a single species, we would ex- 
pect them to grade into each other where their 
distributions are in closest proximity. To examine 
this, we determined the approximate latitude and 
longitude of the collecting locality for each of the 
43 specimens used in the principal components 
analysis of morphology (above) and carried out a 
second principal components analysis on these 
pairs of geographic coordinates. Latitude and lon- 
gitude were evenly weighted (0.707) on the first 
factor axis; hence each score on that first factor 
axis represented the coordinates of one collecting 
locality (Table 5). Plotting the factor scores from 



14 



FIELDIANA: ZOOLOGY 



Table 4. Correlation matrix for the eight variables used in principal components analysis of Cryptotis griseo- 
ventris, C. goodwini goodwini, and C. goodwini magnimana. Abbreviations as in Methods section. 





CBL 


ZP 


IO 


UTR 


MTR 


HCP 


BAC 


ZP 


-0.128 














IO 


0.715 


-0.109 












UTR 


0.234 


-0.173 


-0.004 










MTR 


0.824 


-0.187 


0.641 


0.184 








HCP 


0.798 


0.053 


0.830 


-0.054 


0.738 






BAC 


0.813 


-0.169 


0.784 


0.055 


0.772 


0.774 




mlL 


0.612 


0.132 


0.623 


-0.025 


0.602 


0.652 


0.536 



the first factor axis ("size") of the initial analysis 
of morphological variation against the factor 
scores from the first factor axis of the analysis of 
latitude and longitude ("geographic locality") 
provides a graphical representation of geographic 
variation in size among the three taxa (Fig. 8). We 
calculated regressions of size on geographic lo- 
cality for C. g. goodwini and C. griseoventris to 
aid in the identification of geographic trends in 
size. The slope of the regression line was not sig- 
nificantly different from zero (horizontal) for ei- 
ther C. g. goodwini (y = -1.83 - 0.024*; F = 
0.00, P = 0.956) or C griseoventris (y = 1.72 + 
0.208jc; F = 0.80, P = 0.383), indicating that no 
clear trends in size variation occur within either 
taxon. Therefore, the difference in body size be- 
tween these two parapatric populations is not a 
gradual transition but represents a sharp change. 
One possible explanation might be that some local 



Table 5. Latitude and longitude for collecting sites 
of Cryptotis griseoventris, C. goodwini goodwini, and 
C. goodwini magnimana. "Geographic locality" is a sin- 
gle factor score derived from the first factor axis of a 
principal components analysis of sets of coordinates for 
collecting sites of the three species. Because latitude and 
longitude loaded equally (0.707) on this first factor axis, 
each factor score represents latitude and longitude even- 
ly at a given site. 







Geographic 






locality 


Latitude 


Longitude 


(factor score) 


16°45' 


92°38' 


2.29730 


15°36' 


91°37' 


0.80463 


14°48' 


9i°or 


-0.15638 


15°36' 


91°27' 


0.75103 


14°52' 


91°13' 


-0.05716 


15°45' 


91°29' 


0.84030 


15°04' 


91°34' 


0.50925 


14°45' 


91°33' 


-0.01106 


14°32' 


9o°ir 


-0.77840 


13°40' 


89°29' 


-2.02087 


14°30' 


87°53' 


-2.17865 



selective factor maintains this size difference be- 
tween the two populations. We believe that it is 
much more likely that the difference in body size 
has a genetic basis, and C. g. goodwini and C. 
griseoventris represent two closely related taxa 
that differentiated in isolation and subsequently 
expanded their ranges so that they are now in con- 
tact or near contact. Competition may prevent 
them from coexisting in the same sites. 

Variation in Cryptotis goodwini magnimana is 
impossible to ascertain given a single specimen. 
The holotype of C. g. magnimana is intermediate 
in size between C. g. goodwini and C. griseo- 
ventris, and the population of C. g. magnimana 
probably partially overlaps one or both of the lat- 
ter two taxa in this factor. Cryptotis g. magnimana 
is distinguished morphologically from both C. g. 
goodwini and C. griseoventris by the combination 
of its broader zygomatic plate and its shorter un- 
icuspid toothrow. It is additionally separated from 
C g. goodwini by the less deeply recessed P 4 , M 1 , 
and M 2 (but see below), and from C. griseoventris 
by its absolutely and relatively shorter tail. We 
believe that the morphological distinctiveness of 
the holotype of C. g. magnimana from C. g. good- 
wini probably represents a phylogenetic distinc- 
tiveness as well. Unfortunately, the extremely 
small sample size by which C g. magnimana is 
currently represented prevents us from adequately 
testing this hypothesis. Future study of this prob- 
lem, based on larger sample sizes of the new tax- 
on, will show that it is (1) a species distinct from 
C. goodwini; (2) an individual of C. goodwini that 
shows greater than normal variation for the spe- 
cies and does not warrant a separate name; or (3) 
one extreme of a morphological (not size) cline 
that does not merit a separate name. In support of 
the last hypothesis is the less deeply recessed up- 
per dentition and the short unicuspid toothrow 
(Fig. 9, Tables 1 and 2), two characters that C g. 
magnimana shares with the specimen of C g. 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



15 




geographic locality 



C.g. goodwini: 

O Guatemala 

• El Salvador 

C . g .Tn.ctg7vima.TiCL'. 

□ Honduras 



C. griseoventris: 

V Chiapas 
▼ Guatemala 



Fig. 8. Plot of size against geographic locality for specimens of Cryptotis griseoventris, C. goodwini goodwini, 
and Cryptotis g. magnimana. Size is represented by the scores from the first factor axis of a principal components 
analysis of morphological variation among the three species. Because variables were negatively weighted on the first 
factor axis (Table 3), larger animals are more negative on the y axis, smaller animals more positive. Geographic 
locality is represented by scores from the first factor axis of a principal components analysis of latitude and longitude 
of collecting localities. Localities more to the northwest are more positive on the x axis, those to the southeast are 
more negative. The regression lines for C. g. goodwini (y = -1.83 - 0.024x; F = 0.00, P = 0.956) and C. 
griseoventris (y = 1.72 + 0.208x; F = 0.80, P = 0.383) are not significantly different from zero. 



goodwini from El Salvador. In contrast to C. g. 
magnimana, however, the individual from El Sal- 
vador has an extremely narrow zygomatic plate 
(Fig. 10, Tables 1 and 2). 

Etymology — The subspecific epithet magni- 
mana is an adjectival construction (Latin magnus 
"large" + -i- a connective vowel + manus 



Table 6. Factor loadings for the first three factor 
axes from principal components analysis of Cryptotis 
goodwini from Guatemala and El Salvador. These three 
axes (PCI, PC2, and PC3) together accounted for 77% 
of the variation. Abbreviations as in Methods section. 



Variable 


PCI 


PC2 


PC3 


ZP 


0.272 


0.388 


0.754 


UTR 


-0.204 


-0.493 


0.613 


IO 


-0.324 


0.327 


-0.131 


HCP 


-0.401 


0.463 


0.168 


MTR 


-0.434 


-0.350 


0.014 


M2B 


-0.460 


0.349 


0.003 


CBL 


-0.467 


-0.202 


0.101 



"hand") meaning "large-handed." This is in ref- 
erence to the enlarged forefeet and foreclaws that 
distinguish this taxon from other shrews in Hon- 
duras. The subspecific name takes a feminine ter- 
mination to agree in gender with the feminine ge- 
neric name, Cryptotis (Woodman, 1993). 

Specimen Examined (1)— HONDURAS: CO- 
MAYAGUA: Reserva Biologica Cordillera de 
Montecillos; 2.5 km N, 1.6 km E Cerro San Juan- 
illo [14°30'N, 87°53'W], 1730 m (ku 144611— 
holotype). 



Cryptotis griseoventris Jackson, 1933 

Cryptotis griseoventris Jackson, 1933:80; Hall 

and Kelson, 1959:60. 
Cryptotis goldmani goldmani: Choate, 1970:247 

(in part); Hall, 1981:59 (in part). 

Holotype — Skin and skull of adult male, usnm 



16 



FIELDIANA: ZOOLOGY 



5 














o 

.c 








• 


V 


V 


■*-J 

o 
o 

-t— ' 


2.8 - 




• 


• 
•• • 
V 


V V V w 


V 


■a 
a. 

CO 






•• 


• 
• — 


V V 


V 


D 














O 














c 


2.6 - 






• • • V 


V V V 




D 














O 

r: 




• 






• 


C. griseoventris 


en 


2.4 - 








a 


C. g. magnimana 


c 






a 




C.g.goodwini: 


<D 












V Guatemala 
T El Salvador 




I 




i 


■ 


1 ■ 



19 



20 



21 



22 



23 



condylobasal length 



Fig. 9. Plot of length of unicuspid toothrow (UTR) against condylobasal length (CBL) for Cryptotis goodwini 
joodwini, C. g. magnimana, and C. griseoventris. Both C. g. magnimana and C. g. goodwini from El Salvador have 
datively short unicuspid toothrows. 



CD 

Q. 

U 



o 

en 

N 



o 

CD 













• C. griseoventris 












D C.g. mag nima.no. 


2.2 - 










□ v C.g.goodwini: 
V Guatemala 








• 


• 


• ▼ El Salvador 

V V V 
V 


2.0 - 




• 




• 


# r, A r, V V V 

V • V 












• V V V V V V 


1.8 - 










• • ▼ V V V 
V 




• 










1.6 - 






—r 




• V 

1 1 1 ' 



8.0 



8.5 



9.0 



9.5 



10.0 



length of palate 



Fig. 10. Plot of breadth of zygomatic plate (ZP) against length of palate (PL) for Cryptotis goodwini goodwini, C. 
;. magnimana, and C. griseoventris. Cryptotis g. magnimana has a broad zygomatic plate, whereas C. g. goodwini 
rom El Salvador has a relatively narrow zygomatic plate. 



VOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



17 



6.5 - 



■*-> 
O 

a 






6.0 - 



5.5 - 



# C.alticola 

v C.goldmani 

T C.griseoventris 






• • • 



• • 



vv 



V • V, 

V vv 



V V V V jyT VTV 



19 20 21 

condylobasal length 

Fig. 11. Plot of width of palate (M2B) against condylobasal length (CBL) for Cryptotis alticola, C. goldmani, ant 
C. griseoventris, all of which were considered previously as conspecific. 



75894; taken 4 October 1895 by E. W. Nelson and 
E. A. Goldman (collector number 8545); from 
"San Cristobal" (Jackson, 1933:80) [San Cristo- 
bal de las Casas], Chiapas, Mexico, 9500 ft. 

Distribution — Above at least 2000 m in con- 
terminous highland areas of Guatemala and Chia- 
pas (Fig. 4). 

Description — Size medium for the genus (Ta- 
ble 1); tail moderately long, averaging 29 mm, or 
ca. 38% of HB (Table 2); the dorsal guard hairs 
about 6-8 mm long; dorsal pelage Mummy 
Brown to Bister; ventral fur Light Drab to Drab, 
somewhat paler than dorsum; dorsal fur three- 
banded: basal five-sixths of hairs silvery gray, fol- 
lowed by a thin, lightly colored band grading into 
dark brown at the tip. Rostrum of moderate length 
(PL/CBL = 43.8% ± 0.8, n = 20); usually two 
well-developed dorsal foramina (92%, n = 25); 
no ventral extension of the sinus canal or associ- 
ated foramen (see account for C. goldmani, be- 
low); a foramen dorsal to the dorsal articular facet 
usually present on one or both sides of the skull 
(81%, n = 27); zygomatic plate of moderate 
breadth (ZP/PL = 22.3% ± 1.7, n = 20), anterior 
border usually aligned with mesostyle-metastyle 
valley, metastyle of M 1 , or parastyle of M 2 ; upper 
toothrow uncrowded, U 4 aligned with toothrow 
and partially visible in labial view; P 4 , M 1 , and 
M 2 only very slightly recessed on posterior bor- 



der; M 3 usually with paracrista, paracone, precen 
trocrista, mesostyle, very short postcentrocrist; 
(all pigmented), and a well-developed, lightly pig 
mented protocone. Mandible relatively long am 
of moderate breadth for the genus; articular pro 
cess tall and wide, with a broad lower articula 
facet and a relatively shallow lingual notch be 
tween articular facets; entoconid absent (100%, i 
= 23). 

Comparisons — Cryptotis griseoventris has . 
relatively longer unicuspid toothrow than any o 
the other species except C. g. goodwini. 

Cryptotis alticola — Cryptotis griseoventris ha 
somewhat shorter, narrower foreclaws; a relativel; 
longer tail; lacks any vestige of the ventral exten 
sion of the sinus canal or the associated foramen 
possesses a foramen dorsal to posterior dorsal ai 
ticular facet and two well-developed dorsal fo 
ramina; has a relatively and absolutely narrowe 
palate (Fig. 11); has a relatively and absolutel 
longer unicuspid toothrow; and relatively shorte 
mandible. 

Cryptotis goldmani — Cryptotis griseoventri 
lacks the ventral extension of the sinus canal an i 
associated foramen; possesses a foramen dorsal t 
posterior dorsal articular facet and two well-de 
veloped dorsal foramina; has a relatively and ab 
solutely narrower palate (Fig. 11), relatively an 
absolutely longer unicuspid toothrow, and a lor 



FIELDIANA: ZOOLCKT 



ger, narrower humerus with less enlarged bony 
processes. 

Remarks — Cryptotis griseoventris was origi- 
nally described by Jackson (1933) based on a 
specimen from Chiapas. This was one of many 
lames that Choate (1970) synonymized under the 
subspecies C. goldmani goldmani — an under- 
standable move, given the confused taxonomic 
state of the genus and lack of adequate sample 
sizes at that time. However, a major discontinuity 
;xisted in the range of Choate's C. goldmani gold- 
mani at the Isthmus of Tehuantepec, and popula- 
ions on either side of the isthmus are completely 
isolated from each other. These two populations 
ire easily distinguishable based on a number of 
:ranial and postcranial characters (see Compari- 
sons, above), and it is now clear that they repre- 
sent distinct species. 

Specimens of Cryptotis griseoventris have been 
:aptured above at least 2100 m in forests domi- 
nated by pines and firs and in oak-dominated 
:loud forest. Some areas inhabited by this small- 
jared shrew undergo nightly winter frosts and oc- 
:asional snow (Goldman, 1951). 

No reproductive data are available for Cryptotis 
griseoventris. None of the skins we inspected of 
idult males collected 25-30 September (n = 4), 
2-6 October (n = 2), and 25-31 December (n = 
10) had obvious lateral glands. 

Specimens Examined (28) — GUATEMALA: 
HUEHUETENANGO: Todos Santos Cuchumatan 
[15°36'N, 91°37'W], 10,000 ft (usnm 77051- 
77068). MEXICO: CHIAPAS: San Cristobal de 
las Casas [16°45'N, 92°38'W], 8000-9500 ft 
(usnm 75886-75894, includes holotype); 6 mi SE 
San Cristobal de las Casas (mcz 48061). 



Cryptotis goldmani (Merriam, 1895) 

Blarina mexicana goldmani: Merriam, 1895:25; 

Elliot, 1904:560. 
Blarina mexicana machetes: Merriam, 1895:26; 

Elliot, 1904:561. 
Blarina fossor: Merriam, 1895:28; Elliot, 1904: 

562. 
Cryptotis frontalis: Miller, 1911:222; Hall and 

Kelson, 1959:60; Goodwin, 1969:40. 
Cryptotis mexicana goldmani: Miller, 1912:27; 

Davis and Lukens, 1958:350; Hall and Kelson, 

1959:59; Genoways and Choate, 1967:204. 
Cryptotis mexicana machetes: Miller, 1912:27; 
I Hall and Kelson, 1959:60; Schaldach, 1966: 

288; Goodwin, 1969:40 (in part). 



Cryptotis fossor: Miller, 1912:28; Hall and Kel- 
son, 1959:62; Goodwin, 1969:41. 

Cryptotis guerrerensis: Jackson, 1933:80; Hall 
and Kelson, 1959:60. 

Notiosorex phillipsii: Goodwin, 1969:43 (in part). 

Cryptotis mexicana mexicana: Goodwin, 1969:39 
(in part). 

Cryptotis goldmani goldmani: Choate, 1970:247 
(in part); Hall, 1981:59 (in part). 

Holotype — Skin and skull of adult male, usnm 
70244; taken 23 December 1894 by E. W. Nelson 
and E. A. Goldman (collector number 7231); 
"from mountains near Chilpancingo, Guerrero, 
Mexico" (Merriam, 1895:25), 10,000 ft. 

Distribution — Conterminous high mountain 
areas in the Sierra Madre del Sur and Sierra Ma- 
dre de Oaxaca (Fig. 4); known from above 1500 
m in the Mexican state of Guerrero and above 
2000 m in Oaxaca. 

Description — Size medium for the genus (Ta- 
ble 1); tail moderately long, averaging 29 mm, or 
ca. 38% of HB (Table 2); dorsal guard hairs about 
6-8 mm long; dorsal pelage Mummy Brown; ven- 
tral pelage somewhat paler than that on dorsum, 
between Light Drab and Drab; dorsal fur three- 
banded: basal five-sixths of hairs silvery gray, fol- 
lowed by a thin, lightly colored band grading into 
brown tip. Rostrum of moderate length (PL/CBL 
= 44.2% ± 1.0, n = 16); usually two well-de- 
veloped dorsal foramina (61%, n = 41); an ob- 
vious, well-developed foramen (that opens into a 
ventral extension of the sinus canal) typically is 
present posterior to the dorsal articular facet on 
one or both sides of the skull (97%, n = 38; see 
Remarks, below); a foramen dorsal to the dorsal 
articular facet only occasionally present on one or 
both sides of the skull (16%, n = 38); zygomatic 
plate of moderate breadth (ZP/PL = 21.2% ± 2.4, 
n = 16), anterior border usually aligned with me- 
sostyle-metastyle valley of M', but may extend as 
far posteriorly as the parastyle of M : ; upper tooth- 
row uncrowded, U 4 aligned with toothrow and 
usually partly visible in labial view; P 4 , M 1 , and 
M 2 only very slightly recessed on posterior bor- 
der; M- 1 with paracrista, paracone, precentrocrista, 
and a small, lightly pigmented or unpigmented 
protocone; mesostyle of M 3 is usually small and 
uncolored, and occasionally has a short, uncolored 
postcentrocrista. Mandible of moderate length and 
breadth; articular process tends to be tall and 
wide, with a narrow ventral articular facet and a 
shallow lingual notch between articular facets 
(Oaxaca), or tall and moderately wide, with a 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



19 






Fig. 12. Lateral views of the orbital areas of the skulls 
of A, Cryptotis griseoventris (usnm 75887); and B, 
Cryptotis goldmani (ku 98727). Arrows indicate ap- 
proximate locations of foramen dorsal to (A) and fora- 
men posterior to (B) the dorsal articular facet and as- 
sociated with the ventral extension of the sinus canal. 
Dashed line on B indicates the path of the ventral 
branch of the sinus canal. 



broad ventral articular facet and a deep notch be- 
tween articular facets (Guerrero); entoconid of M 3 
often present, but usually poorly developed (48%, 
n = 23), occasionally well developed (4%). 

Comparisons — Cryptotis goldmani is the only 
member of the C. mexicana-group with an obvi- 
ous, well-developed foramen posterior to the dor- 
sal articular facet. The humerus in this species is 
the shortest and broadest in the genus, and it has 
the most elongate and enlarged bony processes. 

Cryptotis alticola — Cryptotis goldmani aver- 
ages smaller in HB, CBL, and weight (Table 1); 
has a relatively longer tail; has smaller forefeet 
and foreclaws; averages narrower in relative pal- 
atal breadth; and has less recessed posterior bor- 
ders of P 4 , M 1 , and M 2 . 

Remarks — Cryptotis goldmani is unique 
among the broad-clawed Cryptotis in its posses- 
sion of a well-developed foramen associated with 
a ventral branch of the sinus canal (Fig. 12B). 
This ventral branch canal exits through a foramen 
that usually is located posterior to the dorsal ar- 
ticular facet and is approximately the size of the 



anterior orifice of the sinus canal. The foramei 
associated with the ventral branch of the sinus ca 
nal is a convenient character for distinguishin; 
crania of C. goldmani from those of other mem 
bers of the C. mexicana-group. This foramen i 
absent from the holotype of C. g. magnimana anc 
all specimens of C. g. goodwini, C griseoventrh 
and C. nelsoni that we inspected. A minute fora 
men associated with an equally diminutive ventre 
branch canal is present in many C. alticola (549J 
n = 22), C. mexicana (22%, n = 92), C. obscur 
(21%, n = 34), and C. peregrina (85%, n = 20; 
however, the foramina in these species never ap 
proach the size of the well-developed foramen c 
the ventral branch canal present in C. goldman 
Another foramen, dorsal to the dorsal articuk 
facet, often is present among members of the C 
mexicana-group, but this foramen is much les 
prevalent in species (and individuals) that posses 
a ventral branch of the sinus canal, regardless c 
size. 

Cryptotis goldmani, C mexicana, and C. pen 
grina all occur in Oaxaca, and they can be diff 
cult to tell apart, as attested by the number c 
misidentified specimens we have encountere( 
Cryptotis goldmani tends to be larger (HB = 7 
± 5; CBL = 19.6 ± 0.5; see Table 1) than Oa> 
acan C. mexicana (HB = 67 ± 3, n = 16; CB 
= 18.6 ± 0.4, n = 13; C. mexicana from otht 
Mexican states are larger) and C. peregrina (H 
= 66 ± 4, n = 17; CBL = 19.7 ± 0.4, n = 17 
but there is sufficient overlap in size to make th 
an undesirable character for identifying individi 
als with any certainty. Cryptotis goldmani can b 
identified most easily by its much longer ar 
broader foreclaws and, cranially, by its very wel 
developed ventral sinus canal. From C. mexicanc 
C. goldmani usually can be distinguished by i 
vestigial or absent entoconid on M 3 . Cryptot 
mexicana can be distinguished most conclusive 
from C. peregrina by its wider foreclaws and for 
feet and its relatively shorter tail (34% ± 3 of H 
vs. 48% ± 5). Cryptotis mexicana is also muc 
more likely (72%) to have a distinct entoconid < 
M 3 than is C. peregrina (6%). 

Specimens of Cryptotis goldmani have bet 
taken in high-elevation pine forest in Oaxaca ai 
Guerrero. The climates of some localities whe 
the species has been found include severe wint 
frosts and heavy summer rains (Goldman, 1951 
Percy L. Clifton (ku field notes and catalog, 196 
took C. goldmani in snap traps set under rod 
and logs in pine forest with scattered oak trees 
7300 ft at Omiltemi, Guerrero, on 15 July 196 



20 



FIELDIANA: ZOOLOG 



)ther species taken in the same habitat included 
Veotoma mexicana, Peromyscus aztecus, Pero- 
nyscus megalops, and Reithrodontomys sp. Paul 
J. Robertson (ku field notes and catalog, 1970) 
:aptured two C. goldmani while trapping along a 
mall rivulet and a larger stream lined with hard- 
vood trees at 2300 m at Campemento Rio Moli- 
lo, Oaxaca, from 14 to 16 April 1970. The nearby 
lopes were covered with pines. Musser (1964) 
eported that the canyon bottom at Rio Molino 
:ontained relatively moist, open forest with bro- 
neliad-covered oaks and a dense understory of 
hrubs, ferns, and herbaceous vegetation. The 
dgher slopes of the canyon were drier and cov- 
:red with a mixed secondary growth of oaks and 
>ines. In the same trapline, Robertson (ku field 
lotes and catalog, 1970) took Cryptotis peregrina, 
lorex saussurei, Microtus mexicanus, Oryzomys 
'hapmani, Peromyscus megalops, Reithrodonto- 
nys mexicanus, and Rattus sp. In a nearby over- 
grown field, he captured Cryptotis peregrina, Rei- 
hrodontomys megalotis, R. sumichrasti, and Sig- 
nodon alleni. 

Reproductive data for Cryptotis goldmani are 
acking. Lateral glands were well developed on 
hree males collected 21-25 March, three of four 
nales from 18-28 April, three males from 11-15 
une, six males from 25-31 July, and one male 
rom 5 August. A male taken on 26 December 
lad lateral glands that were visible, but not par- 
icularly well developed. Lateral glands were not 
leveloped on single specimens taken during the 
nonths of January, February, and October. The 
iming of gland development that we ascertained 
"or C. goldmani suggests that adult males may be 
eproductively active from March through June, 
iowever, data are lacking entirely for five months 
)f the year, and there are no large series of males 
rom any single month. 

Specimens Examined (58)— MEXICO: GUER- 
RERO: S slope of Cerro Teotepec, 3150 m (ummz 
14710); mountains near [N and NW of] Chil- 
lancingo, 9500-9800 ft (usnm 70243-70245 [in- 
:luding holotype], 70247); Los Retrocesos, 1550 
n (mzfc 3485-3486); El Iris (mzfc 3481-3482); 
)miltemi, 5700-8700 ft, 2450 m (asnhc 3493- 
494; ibunam 29471, 32006; ku 98725-98727; 
acm 74156-74161; mzfc 3480, 3483-3484; 
rsNM 126895-126897 [including holotype of 
:ryptotis guerrerensis], 126947, 127500, 127506); 

mi NW Omiltemi, 2300 m (usnm 329427); 2 mi 

V Omiltemi, 7800-7900 ft (tcwc 5573-5575, 

665); 3 mi W Omiltemi, 8200 ft (mvz 113491); 

; mi SW Filo de Caballo, 8200 ft (tcwc 41948- 



41949). OAXACA: Campamento Rio Molino, 
2300 m (ku 124278, 124301); Vista Hermosa (ku 
143749); Cerro Zempoaltepec, 8000-10,500 ft 
(usnm 68531, 68542, 68545 [holotype of Cryp- 
totis fossor], 68547); Cerro Zempoaltepec, 4.5 km 
N of Santa Maria Yacochi, Mpio. Tlahuitoltepec, 
2450 m (ibunam 33601-33602); Puerto Angel 
Road, lumber camp, km 158, 8375 ft (cas 15477); 
2 km NE San Andres Chichuaxtla, 2300 m (ummz 
113888); Santa Maria Ozolotepec [mountains 
near La Cieneguilla], 10,000 ft (usnm 71454- 
71460, includes holotype of Cryptotis mexicana 
machetes); "near the City of Tehuantepec" (usnm 
123429, holotype of Cryptotis frontalis). 

Additional Records— MEXICO: OAXACA: 
Lachao; San Miguel Suchixtepec; San Juan Ozo- 
lotepec; Mixteguilla (Choate, 1970:249). 



Cryptotis alticola (Merriam, 1895) 

Blarina alticola Merriam 1895:27, Elliot, 1904: 

561. 
Cryptotis alticola Miller: 1912:27; Davis, 1944: 

376; Hooper, 1957:3; Hall and Kelson, 1959: 

60; Genoways and Choate, 1967:204. 
Cryptotis euryrhynchis: Genoways and Choate, 

1967:203. 
Cryptotis goldmani alticola: Choate, 1970:245; 

Hall, 1981:59. 

Holotype — Skin and skull of adult male, usnm 
52047; collected 25 February 1893 by E. W. Nel- 
son (collector number 4396); "from Mount Po- 
pocatepetl" (Merriam, 1895:27) [Volcan Popoca- 
tepetl], Mexico, Mexico, 11,500 ft. 

Distribution — Highland areas above 2000 m 
in the Mexican states of Colima, Jalisco, Michoa- 
can, Mexico, and Puebla, and the Distrito Federal 
(Fig. 4). 

Description — Size medium for the genus 
(Table 1); tail short, averaging 26 mm, or ca. 
33% of HB (Table 2); dorsal guard hairs about 
6-8 mm long; dorsal pelage Olive Brown/Mum- 
my Brown/Bister/Clove Brown; venter some- 
what paler, Light Grayish Olive/Drab/Buffy 
Brown to Olive Brown; dorsal fur distinctly 
three-banded: basal five-sixths of hairs silvery 
gray, followed by a thin, lightly colored band 
grading abruptly to the brownish tip. Rostrum 
of moderate length (PL/CBL = 43.2% ± 1.2, n 
= 16); more than half of specimens examined 
(54%, n = 22) have two well-developed dorsal 
foramina; a minute foramen opening on an 



JVOODMAN & TIMM: 



RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



21 



equally minute, ventral extension of the sinus 
canal may be present posterior to the dorsal ar- 
ticular facet on one or both sides of the skull 
(52%, n = 23; see account for C. goldmani, 
above); a foramen dorsal to the dorsal articular 
facet may be present on one or both sides of the 
skull (39%, n = 23); zygomatic plate of mod- 
erate breadth (ZP/PL = 21.8% ± 1.2, n = 16), 
anterior border aligned with mesostyle-meta- 
style valley of M 1 ; palate wide; P 4 , M\ and M 2 
slightly to moderately recessed on posterior 
border, emargination reaching about halfway to 
hypocone; M 3 with paracrista, paracone, and 
precentrocrista, as well as small, lightly colored 
or uncolored protocone; mesostyle of M 3 usu- 
ally small and uncolored, and sometimes having 
a short, uncolored postcentrocrista. Mandible 
relatively long and of moderate breadth; artic- 
ular process generally tall and moderately wide, 
with a moderately broad lower articular facet; 
entoconid of M 3 often present but very reduced 
(64%, n = 14), or absent. 

Comparisons — Comparisons of Cryptotis alti- 
cola with other broad-clawed shrews are present- 
ed in the above accounts for those species. Se- 
lected measurements are presented in Table 1. 

Remarks — Choate (1970) noted many of the 
differences between Cryptotis alticola and C. 
goldmani (sensu stricto). Although he considered 
the two taxa to be "incipient species," he treated 
C. alticola formally as a subspecies of C. gold- 
mani. We concur with Choate (1970) that it is 
unlikely that there is intergradation between C. 
alticola and C. goldmani. The large number of 
differences in external, cranial, and postcranial 
characters clearly indicates that they are distinct 
species. 

Cryptotis alticola is known to occur in high- 
elevation forests dominated by pines and often 
mixed with firs, oaks, and other tree species. At 
least some of the areas inhabited by C. alticola 
are subjected to winter frosts and occasional 
snow (Goldman, 1951). Percy L. Clifton (ku 
field notes and catalog, 1967) noted that he cap- 
tured three C. alticola along Microtus runways 
on the relatively dry east side of a hill in an 
area of bunch grass under scattered pine, oak, 
and fir at 9000 ft, 20 mi SE of Autlan, Jalisco, 
during 27-28 April 1967. The moister western 
slope of the hill supported denser forest. Other 
small mammals caught in this trapline included 
Sorex saussurei, Microtus mexicanus, Peromys- 
cus aztecus, and Reithrodontomys sumichrasti. 
On 11 June 1967, at 10,000 ft, 12 mi SW of 



Ciudad Guzman, Jalisco, Clifton caught thre 
C. alticola in Museum Special snap traps st 
along a rotten log in a grassy area along th 
edge of a fir, pine, and hemlock forest. Walte 
W. Dalquest (ku field notes, 1946) captured C 
alticola, along with Microtus mexicanus, Nee 
tomodon alstoni, Peromyscus melanotis, an 
Reithrodontomys chrysopsis, in a clearing don 
inated by 3-foot-high clumps of sacaton gras 
in coniferous forest at 11,500 ft, 12 km ESE c 
Amecameca, in the state of Mexico, from 14 t 
15 June 1946. At Monte Rio Frio, Mexico, D; 
vis (1944) captured a C. alticola in an unbaite 
snap trap set along a Microtus mexicanus rui 
way in short grass meadow bordered by pint 
at 10,500 ft. In a thick forest of pines, firs, an 
oaks at 9150 ft, ca. 12 mi W of Ciudad Hidalg. 
Michoacan, Robert W. Dickerman (ku fie] 
notes and catalog, 1955) took C. alticola wit 
Neotomodon alstoni and Reithrodontom) 
megalotis. 

Few reproductive data are available for Cry\ 
totis alticola. A female (ku 17691) from 12 k 
ESE of Amecameca, Mexico, was lactating c 
15 June, and a pregnant female (ku 62311) w; 
found dead along a trail on Cerro Patamba 
Michoacan, on 1 July. Our examination of la 
eral glands on preserved study skins reveals 
pattern similar to that of C. goldmani: later 
glands were well developed on one of tv 
males taken in January, three males captun 
21-25 March, three of four males captured 2' 
28 April, and three males taken in June. Later 
glands were not visible on four males captun 
in February, July, October, and November, r 
spectively. 

Specimens Examined (36)— MEXICO: C( 
LIMA: Volcan de Fuego [= Volcan de Colim; 
8800 ft (lacm 29058). DISTRITO FEDERA 
Cerro de Santa Rosa, 3200 m (ibunam 951 
Santa Rosa, 3000 m (ummz 93367); Canyr 
[Canon] Contreras, 10,200 ft (ummz 9459: 
JALISCO: 20 mi SE Autlan [de Navarro], 9(X' 
ft (ku 1 1 1385-1 1 1387); 12 mi SW Ciudad Gu 
man, 10,000 ft (ku 112039-112041); Volcan J 
Fuego [= Volcan de Colima], 9800 ft (i 
107143, holotype of Cryptotis euryrhynchi.i 
MEXICO: 12 km ESE Amecameca, 11,500 I 
(ku 17691); N edge Refugio San Cayetano, i 
mi S Bosenchere, 8200 ft (ummz 102713); L 
gunas de Zempoala [10 mi NNW Cuernavaca 
Choate, 1970] 9100 ft (usnm 329424-32942* j 
45 km ESE Mexico City, Monte Rio Frio (to 
1927); Estacion Experimental Forestal y de { 



22 



FIELDIANA: ZOOLOC 



Table 7. Complete character matrix for phylogenetic analysis of members of the Cryptotis mexicana-gcoup and 
vo outgroups, Cryptotis parva parva and Cryptotis nigrescens. Transition series are explained in Appendix II. 





























Transitior 


series 
























Taxon 


1 


2 


3 


4 


5 


6 


7 


8 


9 


10 


11 


12 13 14 15 


16 17 


18 19 20 


21 22 23 24 25 26 27 28 29 


jrx'a 























1 





1 



































1 








2 











grescens 














1 
























































1 


1 





1 


1 


exicana 


1 


1 


1 




2 








1 





1 




1 


1 


1 


1 


2 


2 


1 


1 


1 


1 











1 





2 


jlsoni 


1 


1 


1 


9 


2 








1 





1 




9 


9 


9 


9 


9 


9 


9 


9 


9 














1 





2 


bscura 








1 


9 


2 














1 




1 


1 


1 


1 


2 


2 


1 


1 


1 


1 














1 


2 


iregrina 








1 




2 








1 





1 




1 


1 


1 





1 


1 


1 


1 


1 


1 








1 





1 


2 


ticola 


2 


2 


2 




2 











1 







2 


1 


1 


1 


3 


2 


2 


2 


1 


1 








1 


1 





1 2 


jldmani 


2 


2 


2 




2 








1 


1 







2 


1 


1 


3 


3 


2 


2 


3 


1 


2 








1 


1 





1 2 


wdwini 


2 


2 


2 





2 








1 


1 







2 


1 


1 


2 


3 


2 


2 


2 


1 





1 


1 


2 


1 





1 2 


nseoventris 


2 


2 


2 





2 








1 


1 







9 


9 


9 


9 


9 


9 


9 


9 


9 





1 


1 


2 


1 





1 2 


agnimana 


2 


2 


2 


9 


2 








1 


1 







2 


1 


1 


2 


3 


2 


2 


2 


9 





1 


1 


9 


1 





1 2 



lora "L. M. Arellano," 2500 m (ibunam 
4779); Mount Popocatepetl [Volcan Popoca- 
ipetl], 11,500-13,500 ft (usnm 52043, 52045- 
2047, includes holotype); Salazar, 8800- 
0,000 ft (usnm 50757-50759); N slope Volcan 
oluca [= Nevado de Toluca; above San Juan 
e las Huertas], 1 1,500 ft (usnm 55896); 12 km 

San Juan de las Huertas (encb 22675); 15.5 
m S, 7 km E Zinacantepec [= San Miguel Zin- 
:antepec], 3470 m (encb 19357). MICHOA- 
AN: ca. 12 mi W Ciudad Hidalgo, 9150 ft (ku 
6280); 17.5 km NW Ciudad Hidalgo, 2980 m 
•ncb 26210); Sierra Patamba [Cerro Patam- 
an], 9000 ft (ku 66311). PUEBLA: 10 km W 
an Juan Tetla, 3300-3400 m (ibunam 26544- 
6548). 

Additional Records — MEXICO: JALIS- 
CO: N slope Nevado de Colima [= Volcan de 
blima], 8000-10,000 ft. MEXICO: Cerro 
jusco, 11,000 ft. MICHOACAN: Cerro de 
ancitaro [= Pico de Tancitaro], 9000-12,000 

(Choate, 1970:246). 



liylogeny of the Cryptotis 
oldmani-Group 

Our phylogenetic analysis of the Cryptotis 
oldmani- gxonp and the subspecies of Cryptotis 
lexicana, based upon 29 morphological transition 
sries (Table 7), provided nine shortest trees 
ength = 51 steps; Fig. 13). Strict consensus and 
Ldams consensus trees constructed from the nine 
lortest trees are identical (Fig. 14) and represent 
le topology of one of the shortest trees. Areas of 
isagreement among the nine shortest-length trees 
inter on two regions. The first contains the three 



branches leading to C. nelsoni, to C. mexicana, 
and to the Cryptotis goldmani-group. These are 
presented alternatively (1) as an unresolved tri- 
chotomy; (2) with a clade consisting of C nelsoni 
and C. mexicana representing the first outgroup to 
the C. goldmani-group; or (3) with C. mexicana 
as the second outgroup and C. nelsoni as the first 
outgroup to the C. goldmani-group (Fig. 13). The 
second region of disagreement consists of the 
branches leading to C. g. goodwini, C. g. magni- 
mana, and C griseoventris. This region is repre- 
sented (1) as a trichotomy among the three taxa; 
(2) with C. g. goodwini and C. griseoventris as 
sister species; or (3) with C griseoventris and C. 
g. magnimana as sister species. Further resolution 
of these two regions of the trees is hampered by 
the lack of complete skeletal material for C. ob- 
scura and C. g. magnimana, and the lack of any 
skeletons of C. nelsoni or C. griseoventris. De- 
spite this, we have sufficient information to con- 
clusively resolve the most basic relationships 
among these shrews. 

Choate 's (1970) Cryptotis mexicana-group is 
supported by as many as 16 synapomorphies. 
These species all have a similarly shaped skull, 
and they are most easily identified cranially by a 
suite of characters that includes the position of the 
zygomatic plate, the posterior border of which is 
even with (or slightly posterior to), and confluent 
with, the posterior root of the maxillary process 
(Fig. 3). In some North American species of 
Cryptotis, as well as in Blarina and Sorex, which 
commonly are considered to be sister groups of 
Cryptotis (George, 1986), the posterior border of 
the zygomatic plate is anterior to the posterior 
root of the maxillary process. Among members of 
the C. nigrescens complex, the posterior border of 



GOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



23 









S-8SS 



5 S-S 
o'g a Bi 






= ~ £ 
c C 0) 



3 



ilia | 



Q.CfcC(0O)O)O)5oQ. 



_, & 5 2 JS g J 

s §, s * & 5 fi 

Q.C t C «j 0>0)0>t 



is 



5 is 



en 



Ml 



dC t C UOlDlO) 



9 -3 

5 o 



u 



\v 



u 





U 03 



Sill* 8 * 

i: O O C r^ 
03 CD O) CJ> C O Q. 



SS'6 



a I S E 2 




9 <g g 9 o £ -6 

& E g J8 .a 5 o .«) 

to 2> 2 0> *: 5 o c 

ac t c u oioioi 



o 




f to 



• a 'I s 



tj CO 



C C 0) 



3 2 5 £ I 



clgi 



E8| 



to 

? to 

05 C 



c c a> 



03 



?llPi11ilf 

Q.CCCC0O>O>O>COQ. 






g 8i 



tt 



!|l|llilllM 

03-9>20>*:OOC2-O0> 



§ § - 1 2 1 i » 

S (5 P U 5 S i •£ S 

CtC«lC3iO)0)tO 



\^f 



& 




24 



FIELDIANA: ZOOLOC 



he zygomatic plate is even with the posterior root 
jf the maxillary process, as in the C. mexicana- 
*roup, but in the C. nigrescens-group the two 
jtructures are separated by the posterior edge of 
he palate (Woodman & Timm, 1993). Other cran- 
omandibular characters that link members of the 
E mexicana-group are the shape of the coronoid 
jrocess as it joins the horizontal ramus; the rela- 
ively high and narrow articular condyle; and the 
ong, low-cusped P,. These three characters are 
similar to those of Cryptotis gracilis (see Wood- 
nan & Timm, 1992) and may link C. gracilis 
Dhylogenetically with the C. goldmani-group. We 
ire in the process of studying C. gracilis and other 
southern Central American species, and we will 
iddress their relationship to the C. mexicana- 
proup in the future. The members of the Cryptotis 
nexicana- group also show a number of synapo- 
norphies associated with trends toward elonga- 
ion of the foreclaws, broadening of the forefeet, 
ind modification of the humerus. These are con- 
idered further below. 

The four taxa previously considered to be sub- 
pecies of Cryptotis mexicana (C. mexicana, C. 
xelsoni, C obscura, and C. peregrina) do not by 
hemselves form a natural group. The Cryptotis 
loldmani-gxoup is embedded within these taxa, 
ndicating that C. mexicana (sensit lato) is para- 
>hyletic. Among members of the C. mexicana- 
jroup, C. peregrina and C. obscura appear to be 
■he most plesiomorphic species, based primarily 
ipon the development of their foreclaws, their 
mailer body sizes, and their shorter tails relative 
o other members of the C. mexicana-group. 

Monophyly of the Cryptotis goldmani-group is 
upported by at least seven synapomorphies (in- 
luding one reversal). Most of the transition series 
upporting this clade concern the modification of 
tructures of the forelimb. Among these species, 
he two most northern taxa, C. alticola and C. 
oldmani, appear to be the most plesiomorphic 
aembers (but see below). The three taxa from 
outh of the Isthmus of Tehuantepec, C. g. good- 
rini, C. g. magnimana, and C. griseoventris, form 

clade supported by three transition series, all 
ased on cranial foramina. 



Evolution of the Forelimb 

The most distinctive and intriguing character- 
istics of the Cryptotis mexicana-group occur in 
the forelimb. Our studies of these shrews indicate 
that modifications of the forelimb are not restrict- 
ed to the Cryptotis goldmani-group, but represent 
common trends seen throughout the C. mexicana- 
group. Based on our phylogeny, the humerus 
shows a trend of becoming shorter relative to 
overall body size, and the shaft of the humerus 
broadens (Fig. 15). The articular regions become 
enlarged relative to the length of the humerus, and 
various bony processes (notably the teres tubercle, 
medial epicondyle, and pectoral process) elongate 
and enlarge. Due to widening of the base of the 
lateral epicondyle and a lateral shift in the posi- 
tion of the head of the humerus, the humerus ap- 
pears more curved in the mediolateral plane. 
Nearly parallel trends are seen in the enlargement 
of the forefeet (Fig. 16), the shortening and broad- 
ening of the metacarpals and phalanges (Fig. 17), 
and the elongation and broadening of the fore- 
claws. The hindfeet show similar shortening and 
broadening of metatarsals and phalanges (Fig. 
18), but this is not reflected to the same degree 
externally as it is in the forefeet. It is likely that 
some or all of these characters are linked, but we 
did not predict this a priori, and the lack of skel- 
etal material for some taxa makes it impossible to 
confirm the extent of any linkages among these 
traits. For these reasons, we treated forelimb char- 
acters as separate transition series. 

In general, five grades of evolutionary modifi- 
cation of the humerus and the rest of the forelimb 
are exhibited within the Cryptotis mexicana-group 
as compared to the plesiomorphic condition rep- 
resented by the two operational outgroup species, 
Cryptotis parva and Cryptotis nigrescens (Fig. 
15 A). These grades are represented in order by 
(1) C. peregrina (Fig. 15B); (2) C obscura (Fig. 
15C); (3) C. mexicana (Fig. 15D); (4) C. alticola 
and C. goodwini (including C g. goodwini and C. 
g. magnimana; Fig. 15E-G); and (5) C. goldmani 
(Fig. 15H). Within the C. mexicana-group, the hu- 
merus is least modified in C peregrina; the bony 



Fig. 13. Topologies of the nine shortest trees (length = 51 steps; consistency index = 0.82; resolution index = 
86; rescaled consistency index = 0.71) from phylogenetic analysis of the Cryptotis mexicana-group using 29 
lorphological transition series. Cryptotis parva par\'a and Cryptotis nigrescens were used as outgroups. Differences 
mong trees center on two regions: (1) branches leading to C. mexicana, C. nelsoni, and the Cryptotis goldmani- 
roup; and (2) branches leading to C. griseoventris, C. goodwini goodwini, and Cryptotis g. magnimana. 



(VOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



25 



parva 

nigrescens 

mexicana 

nelsoni 

alticola 

goldmani 

goodwini 

griseoventris 

magnimana 

obscura 

peregrina 



Fig. 14. Topology of strict consensus tree of the Cryptotis mexicana-group of species constructed from nine shortej 
length trees. This also is the topology of one of the nine shortest trees and the Adams consensus tree. See Figure 
for additional explanation. 



processes of the humerus are relatively small. Rel- 
ative to the outgroups, however, the humerus of 
C. peregrina is shortened, broadened, and some- 
what curved; the head of the humerus is dorso- 
ventrally elongate; and the distal articular surfaces 
are enlarged (Fig. 15B). The metacarpals of C. 
peregrina (Fig. 17B) are relatively shorter and 
thicker than in C. parva and C. nigrescens (Fig. 
17 A). In the preceding grade, represented by 
Cryptotis obscura (Fig. 15C), the pectoral pro- 
cess, teres tubercle, and medial epicondyle of the 
humerus are enlarged relative to C. peregrina. 
The metacarpals of C. obscura are unknown. Both 
C. peregrina and C. obscura have elongate, but 
narrow, foreclaws and enlarged forefeet (Fig. 
16B). The humerus of Cryptotis mexicana (Fig. 
15D) shows a grade of development similar to 
that of C. obscura. No postcranial material is 
known for Cryptotis nelsoni; however, both C 
mexicana and C. nelsoni have larger forefeet and 
broader foreclaws (Fig. 16C) than C obscura and 
C. peregrina, and the metacarpals of C. mexicana 
(Fig. 17C) are broader than those of C. peregrina. 
Members of the Cryptotis goldmani-group 
have the largest forefeet, the foreclaws are the 
most elongate and broadened (Fig. 16D), and the 
metacarpals and phalanges are the shortest and 
thickest (Fig. 17D). Cryptotis alticola and Cryp- 
totis g. goodwini have the absolutely longest and 
broadest foreclaws, but this may be attributable 
in part to their larger overall body sizes (Table 



1). Postcranial material is unknown for Cryptot 
griseoventris, but C. alticola, C g. goodwih 
and Cryptotis g. magnimana have very simil 
humeri that are relatively shorter and much mo 
curved than in C. mexicana (Fig. 15E-G). Tl! 
bony processes are more elongate, and the ten 
tubercle and medial epicondyle are positiont 
much more closely to each other. The most dl 
rived condition of the humerus is observed | 
Cryptotis goldmani (Fig. 15H). The bony pr< 
cesses are the most pronounced in this specie 
and the teres tubercle and medial epicondyle a 
most elongate and most closely situated to eac 
other, their internal edges forming a symmetric 
reverse C-shape. Cryptotis goldmani was n 
identified as one of the most derived species 
our phylogenetic analysis of the C. mexicam 
group because the characters that most distil 
guish its grade of forelimb development are ai 
tapomorphic. More interesting is the fact th; 
although it has the most derived humerus, < 
goldmani does not possess the most enlarged for 
claws or the broadest forefeet or metacarpals. Th 
suggests that enlargement of forefeet and for 
claws and the change in the shape of the humeri 
are not entirely linked, although the difference 
degree of development of the forefeet and for 
claws may be more closely associated with tl 
body size of the animals involved. 

Small-eared shrews in general have been cor 
sidered to be semi-fossorial or semi-cursori 



26 



FIELDIANA: ZOOLOG 






Fig. 15. Left humeri of selected Cryptotis: A, C. nigrescens (ku 142054, HB not recorded; however, ku 143384 
as a slightly longer humerus and HB = 70 mm); B, C. peregrina (ku 124298, HB = 75 mm); C, C. obscura (mzfc 
45, HB = 65 mm); D, C. mexicana (ku 29541, HB = 78 mm); E, C. alticola (ibunam 26544, HB = 85 mm); F, 
".. g. magnimana (ku 14461 1, HB = 80 mm); G, C. g. goodwini (usnm 275681, HB = 86 mm); and H, C. goldtnani 
ibunam 29471, HB = 84 mm). All humeri are drawn to the same scale. The humeri of C. nelsoni and C. griseoventris 
re unknown. Note the tendency of the humerus to shorten relative to head-and-body length (HB) as the humerus 
ecomes more modified. Scale bar = 1 mm. 



Reed, 1951), although most information about 
lie habits of the genus is derived from studies of 
jst one species, Cryptotis parva, in the United 
tates. This shrew has been observed to be an 
ctive burrower in soft soil, and its nests are often 
Dcated at or below ground level (see Whitaker, 
974, and references therein). However, the small 
;et and foreclaws with which C. parva excavates 
how no obviously specialized adaptations for 
igging relative to other small mammals. Changes 
i forelimb structure that we document for mem- 
ers of the Cryptotis mexicana-group (especially 



the Cryptotis goldmani-group) are distinctive and 
suggest selected specialization, probably for ex- 
cavating. It seems likely that these modifications 
permit this group of shrews to have a more fos- 
sorial niche than other members of the genus. Un- 
fortunately, no behavioral or ecological studies 
are available that might support this hypothesis or 
help us understand the functional significance of 
these forelimb structures. 

Many of the modifications of the humerus 
(shortening, dorsoventral elongation of the head, 
lengthening and enlargement of various processes, 



VOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



27 






Fig. 16. Ventral views of the right forefoot of A, Cryptotis parva orophila (HB = 60 mm); B, C. peregrina (I 
71 mm); C, C. mexicana (HB = 71 mm); and D, C. goldmani (HB = 68 mm). Scale bar = 1 mm. 




B 





Fig. 17. Dorsal views of bones of the left manus of A, Cryptotis parva orophila (ku 142693, HB = 59 mm); 
C. peregrina (ku 124298, HB = 75 mm); C, C. mexicana (ku 29541, HB = 78 mm); and D, C. goldmani (ibun 
33602, HB = 76 mm). The metacarpals of C. p. orophila are nearly the same length as those of C. goldmani, desp 
the large difference in size between the two species. The metacarpals of C. mexicana are slightly shorter and broa< 
than those of C. peregrina. Scale bar = 1 mm. 



28 



FIELDIANA: ZOOLOC 







1 mm 





Fig. 18. Dorsal views of bones of the left pes of A, Cryptotis pana orophila (ku 142693, HB = 59 mm); B, C. 
qrescens (ku 143385, HB = 67 mm); C, C. peregrina'(KU 124298, HB - 75 mm); D, C. mexicana (ku 29541, 
3 = 78 mm); E, C. alticola (ku 62311, no measurements); and F, C. goldmani (ibunam 33602, HB = 76 mm). 
ements with stippled borders are claws. Scale bar = 1 mm. 



OODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



29 



and close association of teres tubercle and medial 
epicondyle), metacarpals (shortening and broad- 
ening), and foreclaws (lengthening and broaden- 
ing) are similar to characteristics seen in the fore- 
limbs of talpids, and they even surpass the degree 
of modification of the humerus observable in 
some of the most primitive grades of moles (no- 
tably Uropsilus). The C. mexicana-group provides 
a living series of structural grades that may pro- 
vide clues as to the early evolution of digging 
adaptations in talpids and possibly other strongly 
fossorial mammals. 



Zoogeography of the Cryptotis 
goldmani-Group 

Choate (1970:297) previously suggested that 
southern Mexico was "the center of origin and 
dispersal" for most modern members of the genus 
Cryptotis because of the large number of species 
there, the high degree of differentiation among 
subspecies, and the presence of what he consid- 
ered to be the most derived species in the genus. 
We concur, based on our understanding of the 
modern biogeography and phylogeny of the Cryp- 
totis mexicana-group, that a highland region of 
central or southern Mexico west of the Isthmus of 
Tehuantepec was the most likely location for the 
evolution of the common ancestor of this clade. 
Most of the eight species of shrews that make up 
the C. mexicana-group are restricted to the west 
of the isthmus, including the least specialized spe- 
cies, Cryptotis peregrina and C obscura. In gen- 
eral, the members of the group inhabit high ele- 
vations; although C. mexicana has been found as 
low as ca. 500 m elevation in Veracruz, elsewhere 
this and all other species in the clade are known 
only from elevations above 1000 m (Choate, 
1970; this paper). 

The distribution of the more specialized Cryp- 
totis goldmani-group is discontinuous from west 
central Mexico to western Honduras. In southern 
Mexico, the occurrence of these shrews corre- 
sponds roughly with the distribution of the tree 
genus Pinus (see Styles, 1993), an observation 
that seems to be supported by habitat descriptions 
in collectors' field notes (see comments on spe- 
cies, above). One of the major gaps in the distri- 
butions of both the Cryptotis goldmani-group and 
Pinus is at the Isthmus of Tehuantepec, where 
there is a break between the highlands of the Si- 
erra Madre del Sur to the west and the Sierra Ma- 



dre de Chiapas to the east. The continental divid 
at the isthmus drops to as low as 250 m, and mon 
tane plant communities give way to vegetatioi 
transitional between lowland rainforest and semi 
arid scrub (Duellman, 1960). Duellman (196C 
1966) recognized the break in the highlands at th< 
Isthmus of Tehuantepec as a major barrier to th« 
dispersal of montane amphibians. In general, th< 
isthmus does not appear to have been regarded a 
a major barrier to most small mammals (Hersh 
kovitz, 1958; Hall, 1981), although it may be im 
plicated in the evolution of subspecies or specie 
of montane Peromyscus (see Carleton, 1989) 
Many middle- to high-elevation mammal species 
including four taxa of soricids {Cryptotis parvi 
pueblensis, C. mexicana, Sorex saussurei, and So 
rex veraepacis), putatively occur on both sides o 
the isthmus (Choate, 1970; Hall, 1981). Howevei 
no species of shrews are known from the isthmu 
itself, and this dry lowland probably is now ai 
effective barrier to the dispersal of all shrews 
Clearly, the isthmus prevents contact betweei 
eastern and western members of the C. goldmani 
group; two species {Cryptotis alticola and C 
goldmani) are found to the west of the isthmus 
and three taxa (C. g. goodwini, C. g. magnimana 
and C. griseoventris) occur on the eastern side. 

Dispersal of the Cryptotis goldmani-grou] 
across the Isthmus of Tehuantepec probably tool 
place during one of the Pleistocene glacial epochs 
as suggested by Duellman (1960, 1966) for mon 
tane amphibians. Climatic amelioration, combine< 
with lowered sea level during one or more of tht 
at least nine Pleistocene glacial maxima, wouh 
have resulted in a lowering or mixing of climati< 
and vegetational zones that permitted montane 
forest plants and animals to extend across the isth 
mus. The subsequent upward migration of cli 
matic zones preceding a subsequent interglacia 
would have isolated populations on either side o 
the isthmus. The sparse paleofloral evidence 
available for southern Mexico and Central Amer 
ica provides support for fluctuation in elevationa 
climatic zones in the past, although regional evi 
dence for the timing of specific vegetationa 
changes remains scant (Graham, 1993). 

The topology of our phylogeny of the Cryptoti. 
goldmani-group (Fig. 14) suggests that thest 
shrews derived from a common ancestor that orig 
inated to the west of the Isthmus of Tehuantepec 
The two species to the west of the isthmus, C 
alticola and C. goldmani, are the two most prim 
itive shrews in the C. goldmani-group in terms o 
numbers of shared-derived characters. Althougl 



30 



FIELDIANA: ZOOLOGY 



C. goldmani is clearly the most specialized mem- 
ber of the C. goldmani-group in terms of its fore- 
limb structure, this degree of specialization is 
mostly autapomorphic. The three eastern taxa (C 
g. goodwini, C. g. magnimana, and C. griseo- 
ventris) form a clade that is supported by up to 
four transition series (TS 22-25 in Appendix I). 
A single character of the humerus (TS 16 in Ap- 
pendix I) links C. goldmani with the three eastern 
taxa, suggesting that they share a common ances- 
tor. Based on our phylogeny and the modern dis- 
tributions of the species in the C. goldmani-group, 
we speculate that the occurrence of members of 
this group on both sides of the Isthmus of Te- 
huantepec is the result of a single dispersal event 
from west to east; as habitat conditions on the 
isthmus became more favorable for shrews during 
a glacial maximum, the common ancestor (of C. 
goldmani and of the three western taxa) expanded 
its range from the west across the isthmus to east. 
As climatic conditions on the isthmus shifted with 
the onset of the succeeding interglacial, the geo- 
graphic range of the ancestral population contract- 
ed into two smaller populations, isolated from 
each other by the isthmus. From the western pop- 
ulation, C. goldmani ultimately evolved, whereas 
the eastern population gave rise to C. goodwini 
and C. griseoventris. (In this scenario, the lineage 
leading to modern C. alticola branched off prior 
to the dispersal across the isthmus.) 

Two lines of evidence provide clues as to the 
timing of the dispersal of the Cryptotis goldmani- 
group across the isthmus. The modern occurrence 
of four species of shrews on both sides of the 
Isthmus of Tehuantepec (but absent from the isth- 
mus itself; see above) suggests that these four spe- 
cies crossed the isthmus during the ultimate (Wis- 
consinan) glacial epoch, but that populations of 
these species on either side of the isthmus, iso- 
lated since the end of that epoch, have had insuf- 
ficient time to differentiate (speciate). (Alterna- 
tively, these "species" may represent unresolved 
taxonomic problems.) Given similar rates of evo- 
lution among these species and the C. goldmani- 
group shrews, we can speculate that the eastern 
and western members of the C. goldmani-group 
separated earlier, possibly subsequent to the pen- 
ultimate (Illinoian) glaciation at the latest. This is 
supported by the recent rediscovery of fossil re- 
mains of C. goodwini-C. griseoventris-grade 
shrews from a site in Honduras believed to be late 
Pleistocene in age (D. Croft, in lit.; Woodman, 
unpublished data). Given an Illinoian or earlier 
dispersal, migrations of the C. goldmani-group 



across the isthmus either to the east or to the west 
during subsequent glacial epochs may have been 
blocked by the presence of sister species (and po- 
tential ecological competitors) in suitable habitat 
on both sides of the isthmus. 



Key to the Cryptotis goldmani-Group 

1. West of the Isthmus of Tehuantepec: usually 
a foramen posterior to the dorsal articular fac- 
et (Fig. 12B); usually lacking a foramen dor- 
sal to external capitular facet; vestigial ento- 
conid may be present onM, 2 

1'. East of the Isthmus of Tehuantepec: always 
lacking a foramen posterior to the dorsal ar- 
ticular facet (Fig. 12A); usually a foramen 
dorsal to dorsal articular facet; entoconid lack- 
ing from M, 3 

2. Larger species (averaging 11 ± 3 g) with 
shorter tail (23-30 mm, averaging 33% of 
HB); foramen posterior to dorsal articular fac- 
et is minute, if present C. alticola 

2'. Smaller species (averaging 8 ± 1 g) with lon- 
ger tail (24-36 mm, averaging 38% of HB); 
large foramen posterior to dorsal articular fac- 
et present C. goldmani 

3. Zygomatic plate broader (ca. 2.2 mm, 25.0% 
of PL); unicuspid toothrow shorter (ca. 2.4 

mm, 1 1.8% of CBL) 

C. goodwini magnimana 

3'. Zygomatic plate narrower (^ 2.2 mm, ^ 
24.7% of PL); unicuspid toothrow longer (^ 
2.5 mm, ca. > 12.0% of CBL) 4 

4. Larger (CBL > 20.0 mm; HB averaging 84 ± 
5 mm) C. goodwini goodwini 

4'. Smaller (CBL < 20.4 mm; HB averaging 77 
± 3 mm) C. griseoventris 



Acknowledgments 

Special thanks to Gustavo Cruz of the Univer- 
sidad Nacional Autonoma de Honduras and to Pe- 
ter Holm of the University of Arizona for making 
important new specimens of Honduran Cryptotis 
available to us for study. We thank the following 
curators and collection managers for loans or for 
permission to examine specimens under their 
care: Guy G. Musser (amnh); Robert C. Dowler 
(asnhc); Luis F. Baptista (cas); Ticul Alvarez and 
Sergio Alvarez (encb); Bruce D. Patterson and 



WOODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



31 



William T. Stanley (fmnh); Fernando Cervantes 
R. (ibunam); Lynn J. Barkley, Sarah B. George, 
and John E. Heyning (lacm); Maria E. Rutzmoser 
(mcz); Elmer C. Birney (mmnh); Livia Leon P. 
(mzfc); Gerhard Storch (smf); George D. Baum- 
gardner (tcwc); Phil Myers (ummz); and Michael 
D. Carleton, Alfred L. Gardner, and Linda K. Gor- 
don (usnm). Robert S. Hoffmann, Guy Musser, 
Ronald H. Pine, Norman A. Slade, and an anon- 
ymous reviewer provided helpful comments on 
previous versions of this manuscript. Amy La- 
throp provided the illustrations used as Figure 6. 
Robert P. Anderson, Rafael L. Joglar, and Adrian 
Nieto Montes de Oca aided with the Spanish 
translation of our abstract. Kate Shaw kindly pro- 
vided technical assistance for our phylogenetic 
analyses. Portions of this project were funded by 
the Panorama Society Fund and the E. Raymond 
Hall Endowment Fund, both of the University of 
Kansas Natural History Museum; the KU Depart- 
ment of Systematics and Ecology, the KU General 
Research Fund; the KU Fellowship Program for 
Latin America Studies; and the National Science 
Foundation (BSR 89-04195); we are indebted to 
each for helping make this study possible. 



Literature Cited 

Bee, J. W., D. Murariu, and R. S. Hoffmann. 1980. 
Histology and histochemistry of specialised integu- 
mentary glands in eight species of North American 
shrews (Mammalia: Insectivora). Travaux du Museum 
d'Histoire Naturelle Grigure Antipa, 22: 547-569. 

Buhler, P. 1964. Zur Gattungs — und Artbestimmung 
von Neomys — Schadeln — gleichzeitig eine Einfiihr- 
ung in die Methodik der optimalen Trennung zweier 
systematischer Einheiten mit Hilfe mehrerer Merk- 
male. Zeitschrift fur Saugetierkunde, 29: 65-93. 

Burt, W. H., and R. A. Stirton. 1961. The mammals 
of El Salvador. University of Michigan Museum of 
Zoology, Miscellaneous Publications, 117: 1-69. 

Carleton, M. D. 1989. Systematics and evolution, pp. 
7-141. In Kirkland, G. L., Jr., and J. N. Layne, eds., 
Advances in the Study of Peromyscus (Rodentia). 
Texas Tech University Press, Lubbock, 367 pp. 

Choate, J. R. 1970. Systematics and zoogeography of 
Middle American shrews of the genus Cryptotis. Uni- 
versity of Kansas Publications, Museum of Natural 
History, 19: 195-317. 

Choate, J. R., and E. D. Fleharty. 1974. Cryptotis 
goodwini. Mammalian Species, 44: 1-3. 

Davis, W. B. 1944. Notes on Mexican mammals. Jour- 
nal of Mammalogy, 25: 370-403. 

Davis, W. B., and P. W. Lukens, Jr. 1958. Mammals 
of the Mexican State of Guerrero, exclusive of Chi- 
roptera and Rodentia. Journal of Mammalogy, 39: 
347-367. 



Duellman, W. E. 1960. A distributional study of th 

amphibians of the Isthmus of Tehuantepec, Mexicc 

University of Kansas Publications, Museum of Natr 

ral History, 13: 21-71. 
. 1966. The Central American herpetofauna: a 

ecological perspective. Copeia, 1966: 700-719. 
Eadie, W. R. 1938. The dermal glands of shrews. Joui 

nal of Mammalogy, 19: 171-174. 
Elliot, D. G. 1904. The land and sea mammals of Mic 

die America and the West Indies. Zoological Serie; 

Field Columbian Museum, 4(2): v-xiii, 441-850. 
Felten, H. 1958. Weitere Saugetiere aus El Salvadc 

(Mammalia: Marsupialia, Insectivora, Primates, Ec 

entata, Lagomorpha, Carnivora und Artiodactyla 

Senckenbergiana Biologica, 39: 213-228. 
Frost, D. R., and D. M. Hillis. 1990. Species in con ; 

cept and practice: Herpetological applications. Hei 

petologia, 46: 87-104. 
Genoways, H. H., and J. R. Choate. 1967. A ne\| 

species of shrew (genus Cryptotis) from Jalisco, Mex j 

ico (Mammalia; Insectivora). Proceedings of the Bi 

ological Society of Washington, 80: 203-206. 
George, S. B. 1986. Evolution and historical biogeog 

raphy of soricine shrews. Systematic Zoology, 35 

153-162. 
Goldman, E. A. 1951. Biological investigations i 

Mexico. Smithsonian Miscellaneous Collections, 115 

i-xiii, 1-476. 
Goodwin, G. G. 1934. Mammals collected by A. V\ 

Anthony in Guatemala, 1924-1928. Bulletin of th 

American Museum of Natural History, 68: 1-60. 
. 1969. Mammals from the State of Oaxac; 

Mexico, in the American Museum of Natural Histon 

Bulletin of the American Museum of Natural Histon 

141: 1-270. 
Graham, A. 1993. Historical factors and biological di 

versity in Mexico, pp. 109-127. In Ramamoorthy, 1 

P., R. Bye, A. Lot, and J. Fa, eds., Biological Diversit 

of Mexico. Oxford University Press, New York, 81 

pp. 
Hall, E. R. 1981. The Mammals of North Amelia 

2nd ed. John Wiley & Sons, New York, 600 + 90 pp 
Hall, E. R., and K. R. Kelson. 1959. The Mammal 

of North America. The Ronald Press Company, Ne\ 

York, 546 + 79 pp. 
Hershkovitz, P. 1958. A geographic classification o 

Neotropical mammals. Fieldiana: Zoology, 36: 581 

620. 
Hooper, E. T 1957. Records of Mexican mammals 

University of Michigan Museum of Zoology, Occa 

sional Papers, 586: 1-9. 
Jackson, H. H. T. 1933. Five new shrews of the genu 

Cryptotis from Mexico and Guatemala. Proceeding 

of the Biological Society of Washington, 46: 79-82 
Maddison, W. P., and D. R. Maddison. 1992. Mac 

Clade: analysis of phylogeny and character evolutior 

version 3. Sinauer Associates, Sunderland, Massa 

chussetts, 939 pp. 
Medellin, R. A. 1988. Prey of Chrotopterus aurituy 

with notes on feeding behavior. Journal of Mammal 

ogy, 69: 841-844. 
Merriam, C. H. 1895. Revision of the shrews of thi 



32 



FIELDIANA: ZOOLOGY 



American genera Blarina and Notiosorex. North 
American Fauna, 10: 5-34. 

Miller, G. S., Jr. 191 1. Three new shrews of the genus 
Cryptotis. Proceedings of the Biological Society of 
Washington, 24: 221-224. 

. 1912. List of North American land mammals 

in the United States National Museum, 1911. Bulletin 
of the U.S. National Museum, 79: 1-455. 

Murariu, D. 1976. Les glandes tegumentaires de cer- 
tains insectivores (Mammalia — Insectivora) de Rou- 
manie. Anatomie, histologic et histochimie. Travaux 
du Museum d'Histoire Naturelle Grigure Antipa, 17: 
387-413. 

Musser, G. G. 1964. Notes on geographic distribution, 
habitat, and taxonomy of some Mexican mammals. 
Occasional Papers of the Museum of Zoology, Uni- 
versity of Michigan, 636: 1-22. 

Ieed, C. A. 1951. Locomotion and appendicular anat- 
omy in three soricoid insectivores. American Midland 
Naturalist, 45: 513-671. 

Iidgway, R. 1912. Color standards and color nomen- 
clature. Published privately by the author, Washington, 
D.C., 43 pp. + 53 pi. 

Jchaldach, W J., Jr. 1966. New forms of mammals 
from Southern Oaxaca, Mexico, with notes on some 
mammals of the Coastal range. Saugetierkundliche 
Mitteilungen, 14: 286-297. 

Jtuart, L. C. 1954. A description of a subhumid cor- 
ridor across northern Central America, with comments 
on its herpetofaunal indicators. Contributions from the 
Laboratory of Vertebrate Biology, University of Mich- 
igan, 65: 1-26. 

Ityles, B. T. 1993. Genus Pinus: A Mexican purview, 
pp. 397-420. In Ramamoorthy, T. P., R. Bye, A. Lot, 
and J. Fa, eds., Biological Diversity of Mexico. Ox- 
ford University Press, New York, 812 pp. 

iwOFFORD, D. L. 1993. PAUP: Phylogenetic analysis 
using parsimony, version 3.1. Illinois State Natural 
History Survey, Champaign. 

J.S. Board on Geographic Names. 1956a. Gazetteer 
no. 15. Mexico. U.S. Government Printing Office, 
Washington, D.C. 

. 1956b. Gazetteer no. 26. El Salvador. U.S. 

Government Printing Office, Washington, D.C. 
. 1965. NIS Gazetteer. Guatemala. U.S. Govern- 



Soricidae), from Honduras. Proceedings of the Bio- 
logical Society of Washington, 105: 1-12. 

1993. Intraspecific and interspecific variation 



ment Printing Office, Washington, D.C. 

/hitaker, J. O., Jr. 1974. Cryptotis par\>a. Mamma- 
lian Species, 43: 1-8. 

/iley, E. O. 1978. The evolutionary species concept 
reconsidered. Systematic Zoology, 27: 17-26. 

1981. Phylogenetics. John Wiley & Sons, New 



York, 439 pp. 

I J/oodman, N. 1992. Biogeographical and evolutionary 
relationships among Central American small-eared 
shrews of the genus Cryptotis (Mammalia: Insectivo- 
ra: Soricidae). Ph.D. diss., University of Kansas, 386 
pp. 

. 1993. The correct gender of mammalian ge- 
neric names ending in -otis. Journal of Mammalogy, 
74: 544-546. 

Goodman, N., and R. M. Timm. 1992. A new species 
of small-eared shrew, genus Cryptotis (Insectivora: 



in the Cryptotis nigrescens species complex of small- 
eared shrews (Insectivora: Soricidae), with the de- 
scription of a new species from Colombia. Fieldiana: 
Zoology, new series, 74: 1-30. 



Appendix I: Transition Series Used in 
Phylogenetic Analysis 

1. length of foreclaws (Fig. 16): short (0); elon- 
gate (1); greatly elongate (2). 

2. breadth of foreclaws (Fig. 16): narrow (0); 
broadened (1); greatly broadened (2). 

3. forefeet (Fig. 16): small (0); enlarged, broad- 
ened (1); greatly enlarged and broadened (2). 

4. metacarpals (Fig. 17): long, narrow (0); short, 
broad (1). 

5. posterior border of zygomatic plate (Fig. 3): 
even with or anterior to anterior root of max- 
illary process (0); even with posterior root of 
zygomatic process, but separated from it by 
posterior border of palate (1); even with (or 
posterior to) and confluent with posterior root 
of zygomatic process (2). 

6. anterior border of coronoid process: steep, 
forming a narrow angle with horizontal ramus 
of mandible (0); less steep, forming a wide 
angle with horizontal ramus of mandible (1). 

7. articular condyle: low and broad (0); high and 
narrow (1). 

8. lower sigmoid notch: very shallow (0); deep 

(1). 

9. shape of unicuspids (U'-IP): cone-shaped, 
posteroventral border straight-edged or con- 
vex (0); narrow, posteroventral border con- 
cave (1). 

10. protoconal basin of M': about equal in size to 
hypoconal basin (0); reduced relative to hy- 
poconal basin (1). 

11. M 3 morphology: simple, metacone absent (0); 
complex, metacone present (1). 

12. shape of P 3 : short and high (0); long and low 

(D- 

13. shape of humerus (Fig. 15): long, narrow, rel- 
atively straight (0); short, robust, curved (1); 
short, robust, and very curved (2). 

14. head of humerus: rounded (0); dorsoventrally 
elongate (1). 

15. ventral edge of proximal face of greater tu- 
berosity of humerus: rounded (0); with broad, 
deep pocket (1). 



/OODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



33 



16. pectoral process of humerus: low (0); high 
(1); very high (2); extremely high (3). 

17. teres tubercle of humerus (Fig. 15): relatively 
short (0); elongate (1); moderately elongate 
(2); greatly elongate (3). 

18. medial epicondyle of humerus (Fig. 15): short 
(0); elongate (1); greatly elongate (2). 

19. lateral epicondyle of humerus (Fig. 15): small 
(0); expanded (1); greatly expanded (2). 

20. teres tubercle and medial epicondyle of hu- 
merus (Fig. 15): far apart (0); close together 
(1); very close (2); extremely close (3). 

21. posterior edge of falciform process of tibia: 
not deeply pocketed (0); deeply pocketed (1). 

22. foramen of sinus canal (Fig. 12): absent (0); 
present, but tiny (1); well developed (2). 

23. dorsal foramina: two in < 75% of specimens 
(0); two in > 75% of specimens (1). 

24. foramen dorsal to articular facet (Fig. 12): 
present in < 75% of specimens (0); present 
in > 75% of specimens (1). 

25. entoconid of M 3 : well developed, present in 

> 75% of specimens (0); vestigial, but pres- 
ent in < 76% of specimens (1); absent (2). 

26. body size (head-and-body length): smaller, 
mean < 69 (0); larger, mean > 72 (1). 

27. relative tail length (% of head-and-body 
length): short, mean < 39% (0); long, mean 

> 40% (1). 

28. upper unicuspid toothrow: crowded, three un- 
icuspids visible in lateral view (0); uncrowd- 
ed, four unicuspids visible in lateral view (1). 

29. dorsal guard hairs: short, < 4 mm (0); of in- 
termediate length, 4-5 mm (1); long, > 5 mm 
(2). 



Appendix II: Additional Specimens 
Examined 

Cryptotis mexicana (59)— MEXICO: CHIA- 
PAS: 3 mi E Pueblo Nuevo Solistahuacan, 7000 
ft (ku 83942); OAXACA: Vista Hermosa, 1500 
m (ku 91464); 6.5 mi SSW Vista Hermosa, 7100 
ft (ku 99546); 21.8 km S Vista Hermosa, 2100 m 
(ku 124274); 31.6 km S Vista Hermosa, N slope 
Cerro Pelon, 2650 m (ku 124275-124276); NE 
slope Cerro Pelon, 2620 m (ku 120302); Cerro 
San Felipe, 6 km W La Cumbre, 2670 m (ku 
121658, 124293); 2 km W La Cumbre, 2900 m 
(ku 121414); 7 mi N Ixtlan de Juarez, 10,000 ft 
(cas 12244-12245); Llano de las Flores, 2800- 



3150 m (ku 91465, 120303, 121410-12141 
124277); 0.4 mi S Llano de las Flores, 9200 
(tcwc 45106); 11 mi NE (Tuxtepec road) Liar 
de las Flores, 9100 ft (ummz 112571); 27.5 k 
NNE Llano de las Flores (lacm 74166-74173 
Ixtlan, 16 mi WSW La Esperanza (tcwc 2501 1 
PUEBLA: 7.5 mi NE Tezuitlan (mmnh 6875); 
mi NE Tezuitlan (mmnh 4710, 6939). VERV 
CRUZ: Las Vigas, 8500 ft (ku 29525, 2952* 
29540, 29542-29545, 29547-29549); 11 km ^ 
Las Vigas, 8500 ft (ku 29524); 5 mi E Las Vig, 
(tcwc 25077); 4 km W Tlapacoyan, 1700 ft (v 
23412-23414). 

Cryptotis nelsoni (9)— MEXICO: VER/ 
CRUZ: Volcan Tuxtla, 4800 ft (mcz 19747; usn 
65429-65433, 65435-65437, includes holotype 

Cryptotis nigrescens (62) — COSTA RIC/ 
ALAJUELA: Cinchona, 1600 m (ku 106942 
Monteverde Cloud Forest Reserve, 1580-1600 I 
(ku 143377-143380); Monteverde Cloud Fore] 
Reserve, Pefias Blancas Valley, 870 m (v j 
143371, 143381). CARTAGO: [Volcan] Iraz j 
8000-9400 ft (amnh 141200). GU AN AC AST 
PUNTARENAS border: Monteverde, Cerro Arr 
gos, 1790 m (ku 143382, 143384). PUNTA1 
ENAS: Coto Brus [Canton], Sabalito District, L 
Tablas, Rio Coton, 1700 m (mncr no numbej 
Monteverde, 1345-1600 m (fmnh 12410 
128415, 135224; iNBio no number; ku 13489 
135008, 135083, 142053, 142689, 14278* 
142789, 143295-143297, 143383, 14338: 
143395, 143636, 143638, 144612; lacm 6484 
67443, 67453; mmnh 14095; ummz 11588.' 
115884, 117107-117110, 125632); Monteverd 
Cerro Amigos, 1760 m (ku 142054); Monteven 
Cloud Forest Reserve, 1530-1660 m (ku 14278( 
142787, 143386, 143396, 143496, 143637); S; 
Luis, 1200 m [ca. 2.5 km S Monteverde] (t 
143385); 1 mi SW of Finca Las Cruces, San Vit 
4000 ft (lacm 74351-74353). SAN JOSE: S; 
Isidro (amnh 7952/9691, holotype); 9 mi N of S; j 
Isidro del General (Pan American Highway), 48( I 
ft (ummz 111999). 

Cryptotis obscura (42)— MEXICO: HIDAJ 
GO: Tlanchinol, 20°59'N, 98°39'W (ibuna 
4186); Tulancingo (usnm 55633); Encarnacic 
(usnm 81125-81127, 81131-81134); 11 km E 
Acaxochitlan, Lago Tejocotal, 2250 m (]• 
81770). QUERETARO: 1 km S Ahuacatlan (mz 
637); Pinal de Amoles (usnm 81115-8111 
81123-81124, 81129, 81147); 1 km S Pinal i 
Amoles (mzfc 638-648); 4 km SW Pinal de Am 
les, 2550 m (ibunam 29106, 29332). TAMA1 
LIPAS: 5 mi NW of Gomez Farias, Rancho d 



34 



FIELDIANA: ZOOLOG 



:ielo, 3500 ft (mmnh 4301-4305, 4570-4571, 
■74). VERACRUZ: Zacualpan, 6000 ft (ku 
58241). 

Cryptotis parva orophila (25) — COSTA 
RICA: ALAJUELA: Santa Clara (mncr no num- 
>er). ALAJUELA: Zarcero, 6000 ft (fmnh 43974). 
ZARTAGO: Cartago (ku 26932; ummz 66465, 
37316); Coliblanco (ku 26930-26931); La Estrel- 
la (amnh 14847); Estrella de Cartago, 4500 ft 
ummz 64147); Guarco (ku 16563); Navarro (mcz 
>1656); Irazu Range [Volcan de Irazu] (amnh 
>64 1/9841, holotype). HEREDIA: Barva Canton, 
5an Jose de la Montana, Paso Llano, 1800 m (ku 
142692-142694); Barva Canton, San Miguel de 
a Montana, 1690-1700 m (ku 143372-143374). 
SAN JOSE: 10 mi S of Cartago, El Muneco, 3800 
t (ummz 67315); Cerro Tablazo, 1983 m (usnm 
52525); San Jose (usnm 7224, 38477); San Pedro 
le Montes de Oca (amnh 139282); San Rafael de 
Pontes de Oca, 4300 ft (ku 147100); Santa Ana 
ilsu 15753). 



Cryptotis parva parva (48)— USA: KANSAS: 
Douglas Co.: Lawrence (ku 22131-22142); 3.5 
mi W of Lawrence (ku 125554-125568); 1.5 mi 
N, 1.7 mi E of Lawrence (courthouse), sandpits 
(ku 114236-114239, 114241-114247). 

Cryptotis peregrina (18)— MEXICO: OAXA- 
CA: mountains 15 mi SW Oaxaca de Juarez, 9500 
ft (usnm 68317, holotype); Rio Molino, Puerto 
Angel road, km 153, 2250-2300 m (cas 14068; 
ku 121661; ibunam 8447); lumber camp, Puerto 
Angel road, km 158, 8375 ft (cas 15478); Rio 
Jalatengo, Puerto Angel road, km 178, 4275 ft 
(cas 14069, 14071-14072, 15475; ibunam 
27518); Puerto Angel road, km 195, 3475 ft 
(ibunam 26551); La Cima, Puerto Escondido road, 
km 184.5, 5750 ft (cas 15473); Puerto Escondido 
road, km 193, 4200 ft (cas 15474); Rio Guajalote, 
2000 m (ku 114226); Sinai, 10 km E Nopala, 
7200 ft (cas 14940); 20 mi S, 5 mi E Sola de 
Vega, 4800 ft (ku 98728); 16 km SW Suchixte- 
pec, 2000 m (encb 3413-3414). 



OODMAN & TIMM: RELATIONSHIPS AMONG BROAD-CLAWED SHREWS 



35